UNITED NATIONS CONFERENCE ON TRADE AND DEVELOPMENT
20 April 2005
ENGLISH ONLY
SUBSIDIES TO SERVICES SECTORS: A NEO-PROTECTIONIST DISTORTION
OR A USEFUL DEVELOPMENT TOOL? 1
UNCTAD/DITC/TNCD/MISC/2003/7
1
The report has been prepared by Mr. Alberto Gabriele, Economic Affairs Officer, Division on International
Goods and Services, and Commodities, UNCTAD.
INTRODUCTION
The purpose of this paper is twofold. The first objective is analytical and the second is policyoriented. The first section of the paper illustrates a few basic concepts on subsidies; the
second assesses the main costs and benefits and the institutional feasibility of implementing
subsidy programmes from a political economy viewpoint, focusing chiefly on services
sectors; the third examines the evolution and outcomes of the controversies on subsidies in
the GATT/WTO system of multilateral trade rules; and the fourth presents a synthetic
quantitative picture of the pervasiveness and magnitude of subsidy policies worldwide,
stressing their unequal weight and impact in developed and developing countries
respectively. The fifth section provides some elements contributing to the elaboration of an
understanding of the risks and opportunities that subsidy policies present for developing
countries, and to the development of a common negotiating stance in forthcoming WTO
negotiations on subsidies. These sections are described in more detail below.
Concepts and definitions
This section illustrates the basic concept of subsidy and provides a few canonic definitions
from the viewpoint of mainstream economic theory, referring to issues such as: the main
justifications put forward to implement subsidy policies; the relationship between subsidies,
Pareto efficiency and the inter-temporal allocation of resources; subsidies as policy tools to
overcome market failures caused by imperfect and asymmetric information; subsides as
social, distributional and environmental policy tools; and the trade implications of subsidies,
or subsidies as trade policy tools.
The political economy of subsidizing services
This section considers the advantages and disadvantages of subsidy programmes. It pays
particular attention to the possible virtuous role of subsidies in development strategies.
Subsection 2.1 discusses why many subsidy programmes, as well as other interventionist
policies, are implemented and why they tend to show remarkable persistence over time, in
spite of the fact that many observers routinely argue against their inefficient, perverse and
ultimately irrational consequences. Subsection 2.2 explores the likelihood of cooperative and
conflictive outcomes in international trade policy rivalries (an issue also investigated in
Box 2.2). Subsection 2.3 focuses on the role of services as strategic trade policy tools, with
particular attention paid to education and R&D services. Boxes 2.3 and 2.4 refer to the impact
of subsidies on international trade in two other important service sectors, namely construction
and health, taking stock of the debate that took place in two UNCTAD expert meetings.
Subsidies in developed and developing countries: a review of quantitative evidence
This section discusses the complexity of the task of evaluating global subsidies worldwide in
a comprehensive and exhaustive fashion, and presents some synthetic figures indicating the
order of magnitude of total subsidy programmes worldwide. Subsections 3.3 to 3.5 report
estimates of expenditures on subsidies in the areas of water, energy and transport services in
developed and developing countries. Subsection 3.6 focuses on subsidies in the fields of
education and R&D.
Services in multilateral trade negotiations
This section briefly recounts the history of the subsidy issue in the framework of multilateral
negotiations, describes the treatment of subsidies in GATT and GATS and the main features
2
of the Agreement on Subsidies and Countervailing Measures, and reviews some documents
and analytical contributions put forward in the Working Party on GATS Rules. The relative
resilience of subsidies and other indirect trade policy instruments, in contrast to the long-run
decline of traditional protectionist tools, is also highlighted. In the final part of the section,
three boxes focus on specific aspects of themes examined in the main text. Box 4.1 proposes
an interpretation of the law of subsidies centred on the principle of attenuating the entitlement
of affected parts as a precautionary device aimed at avoiding major disruptions in
international trade flows. Box 4.2 describes the issue of the definition of R&D subsidies in
multilateral trade discipline and on the evolution of the US negotiating position. Box 4.3
stresses the particularly wide policy space still available to Governments willing to
implement subsidy policies to support audiovisual services.
Subsidies in the area of services in developing countries: a necessary policy tool for
developing countries in an asymmetric world
This short concluding section summarizes the finding of the previous four in order to assess
the role and feasibility of subsidies as a development strategy tool in developing countries.
The importance of proper targeting, planning, focusing and timing of subsidy programs is
underlined, as is the weight of the objective financial constraints that inevitably put less
developed countries at a relative disadvantage in the event of the eruption of a subsidy-based
trade war. The main conclusion is twofold. On the one hand, developing countries should
strive to achieve a widely shared and possibly common position in WTO negotiations on
subsidies, trying to pose the effective recognition of special and differential treatment in the
domain of subsidies as a necessary condition for pursuing further the agenda on progressive
liberalization. On the other hand, the international community at large, and developed
countries in particular, should recognize the acute asymmetry of conditions among countries
at very different levels of development. This asymmetry makes it imperative not only to
reconfirm the spirit of existing WTO rules on subsidies, but also to substantiate them to a
much greater extent.
1.
CONCEPTS AND DEFINITIONS
1.1.
Justifications for subsidies under a simplified closed economy assumption
Subsidies, like taxes, are among the most common policy tools used by Governments to get
the economy to pursue a variety of economic, social or environmental objectives. There can
be many different justifications for subsidies, depending on the specific policy goals they are
meant to pursue and on the scenario for their implementation. In this respect, even if it is only
a working approximation (as no fully autarkic economy exists in the present, globalizing
world), it is useful to refer initially to the traditional closed economy assumption, where the
Government pursues exclusively non-trade, domestic policy goals. The main traditional
justification for subsidies (taxes), according to traditional mainstream economics, is as an
instrument (even if a second-best one) to correct market failures, thus helping the economy to
achieve a more efficient equilibrium than the one that would prevail in a laissez-faire
situation.
Market failures can manifest themselves as positive or negative externalities, and can be seen
in a static or a dynamic dimension. From a development strategy point of view, however, the
dynamic dimension is of particular interest. The market, left to function spontaneously, is
bound to be subject to inter-temporal failures: such failures essentially arise because in
economic terms individual agents tend to behave according to a subjective discount rate
3
which is different – usually too high – in relation to the optimum discount rate a hypothetical
perfectly informed planner would assign to the society as a whole. In other words, individual
agents often exhibit short-sighted behaviour, investing too little and shunning complex,
uncertain long-term projects, partly but not exclusively due to lack of information. On the
contrary, policy makers do have (or should have) a strategic development vision. Thus, they
can implement a development strategy, earmarking (through subsidies and other instruments)
resources for new and strategic industries. Subsidies aimed at protecting the environment can
also be seen as measures to achieve efficiency in a broad and long-term (inter-generational)
dimension, internalizing external diseconomies that are not accounted for by the market and
pushing the economy towards a more sustainable development path.
Inside a particular industry, or horizontally via measures affecting all sectors of the economy,
Governments can also prioritize specific subsectors or activities that are deemed to have a
particularly high potential to boost productivity growth and technological progress across the
national economy as a whole. Most of these activities are of an infrastructural or, more
broadly, of a service character, and thus the pros and cons of their subsidization are especially
relevant in the context of this paper, taking into account its sectoral focus.
In this respect, a particularly important category of subsidies, often justified on the basis of
both static and dynamic externalities, is that of subsidies for R&D. In a closed economy
context, these subsidies are often justified on the grounds that private firms left on their own
would under-invest in R&D, as the corresponding benefits appear too subject to risk and too
far into the future from the point of view of the single economic agent. The issue of
subsidizing R&D becomes more complex in the more realistic context of an open economy
(Box 1.1).
Another important category is constituted by subsidies targeted at achieving social
distributional goals: subsidies to the poor, the unemployed, handicapped people, less
developed regions, and the like. There can be cases when such subsidies are fully Paretoefficient, but usually they are not: for example, subsidies to the unemployed might stifle the
willingness to work, and subsidies to disadvantaged regions might distort incentives and
cause a less than fully efficient nationwide allocation of resources. In these cases policy
makers must assume the existence of a trade-off between efficiency and a social goal, to be
resolved according to the relative weight attached to each. From this viewpoint, too, subsidies
to some service activities are exceptionally important. Subsidized – usually public –
provision of essential services, such as security, education and health, and possibly of many
other services needed for the direct fulfilment of crucial welfare requirements (such as care
for children, the elderly and disabled, transportation, and cultural services), is an essential
policy tool aimed at ensuring the broadest possible satisfaction of basic human needs and, at
the same time, equalizing the real distribution of income.2
2
As an instrument to improve the real distribution of income, the direct provision of basic services can often be
preferred to other policy tools, as it does not affect the distribution of private monetary incomes and therefore
might be deemed to be less distortionary.
4
Box 1.1
Subsidies in an open economy: a textbook model
The basic (static) functioning of subsidies in an open economy, according to mainstream theory, is
illustrated in most manuals of international economics and trade. Among them, Gandolfo (1998) proposes a
particularly clear exposition, based on a static, partial equilibrium analysis approach, which can be particularly
useful in evaluating the welfare cost of alternative subsidy policies.
The subsidy policy might be well justified, according to priority criteria that cannot be captured by this
simple model, as it could be needed to improve macroeconomic equilibrium (i.e. the balance of payments).
Alternatively, or in a complementary fashion, the subsidy could generate long-term dynamic benefits, providing
national entrepreneurs with additional incentives to channel more resources towards technologically advanced,
fast-growing products. However, according to mainstream theory, the subsidy also inevitably implies a static
welfare loss, which is constituted by the sum of two components. One component is the welfare loss in terms of
consumers’ surplus, and the other the loss in terms of producers’ surplus. Both losses are caused by the
distortions implied by the shift in equilibrium caused by the subsidy, taking into account consumers’ and
producers’ supply and demand function.
Subsidies can be earmarked for both a (net) exporting sector and a (net) importing sector. In the former
case, “the subsidy can be either an export subsidy (i.e. given to domestic producers on the exported part of their
output) or a production subsidy (i.e. given to domestic producers on their whole output” (Gandolfo 1998, p.
209).
If the subsidy is an export subsidy, the supply schedule does not change: producers move outwards
along the curve, increasing export-bound production to make the most of the subsidy. National producers
prioritize overseas markets, and the (residual) domestic consumption shrinks, causing a decrease in consumers’
surplus. The total welfare cost is given by the sum of the welfare costs from the consumers’ and the producers’
side respectively, caused by the “distortions” induced by the subsidy. The Government, alternatively, can
choose to introduce a production subsidy. The supply function shifts downward, and equilibrium is reached at a
higher level of output. All additional output is exported, and the domestic price does not change. In comparison
with the export subsidy, the financial cost for the Government (ceteris paribus) is higher (all production, and not
only its export share, is now subsidized), but the welfare cost is lower (as the domestic price does not change,
domestic consumption does not change either and thus there is no loss of welfare on the consumers’ side).
The conclusion from this (admittedly narrow) partial equilibrium approach is therefore that a
production subsidy, which does not create a wedge between the domestic and the international price, is
preferable to direct trade intervention, such as an explicit export subsidy. From the point of view of practical
political feasibility, moreover, the potential advantage of a production subsidy is even more evident: as opposed
to an export subsidy, it is less likely to be unambiguously forbidden by multilateral trade rules. A production
subsidy (especially if finely crafted as an indirect subsidy to one or more production inputs, i.e. energy, R&D,
transport, or others) is also much easier to portray as a policy measure mainly targeted at non-trade goals, often
not intrinsically economic in nature (i.e. social, environmental), rather than at increasing exports. A third
scenario contemplates a net importing subsidy, in which the Government grants domestic producers a
(production) subsidy. As a result, the supply schedule shifts upwards, national production increases and imports
shrink. It is interesting to note that in this case – although, as in the preceding one, the subsidy is spread over the
totality of domestic production – market conditions are different. National consumers keep purchasing the same
amount at the same (international) price; the only change is that now the share of domestic producers has
increased. Therefore there is no loss of consumer surplus. This outcome (again, keeping in mind the caveat that
the analytical approach is a very partial and stylised one) shows that, if the Government’s goal is to reduce
imports, a producer subsidy can be a Pareto-superior policy tool with respect to a tariff (which distorts both
production and consumption, implying welfare losses on both sides).
Summing up the results of the simple partial equilibrium models outlined above, Gandolfo (1998,
p.211), expressing a view shared by most mainstream economists, concludes that “production subsidies are to be
ranked above tariffs” . That the matter is not only one of textbook theoretical disputes is made clear immediately
below, where the author observes that “the previous treatment explains why countries may prefer the
instruments of the new protectionism rather than the traditional ones” (ibid.).
5
Arguments in favour of subsidies must be weighed against those militating against, which are
often focused on the negative dead-weight effect they may cause to the economy due to their
distortionary effects. Besides that, Governments (in good or bad faith) might end up unduly
favouring some firms or social groups at the expenses of others, and might in fact lack the
information to “pick the winners” among potential national champions. Moreover, even if
possibly to a lesser extent than in the case of other protectionist tools, the abuse of subsidies
could become self-reinforcing and ultimately impossible to keep in check, leading not only to
chronic inefficiencies but also to the risk of catastrophic trade wars.
1.2.
Subsidies in an open economy engaged in international trade
It would be extremely difficult to evaluate fully, in a general equilibrium framework, the
magnitude of all direct and indirect benefits and costs induced by subsidy policies worldwide,
or their distribution among different social groups and geographic areas
(subsection 1.4). However, it is safe to assume that the vast majority of subsidy policies are
enacted mostly to pursue domestic goals, and most of their impact is also felt domestically
(although the indirect international effect of many subsidy policies implemented chiefly for
domestic purposes can also be quite large, as is for instance the case with agricultural support
policies).
Yet there is another category of subsidies which, even if its budgetary weight is relatively
modest with respect to subsidies aimed at domestic policy goals, has a disproportionate
impact on countries’ international competitiveness, and thus on worldwide trade flows. These
are the subsidies specifically aimed at achieving trade policy goals. In this perspective, it is
important to note that policy makers do not always implement subsidy programmes and other
proactive interventions with the main purpose of favouring special minority interest groups at
the expense of the rest of the national population. Rather, their goal is to strategically
strengthen their own country (admittedly, frequently identified rather narrowly with the
business, industrial and political elite) at the expense of other countries. This is the essence of
strategic trade policies, which rely on subsidies as one of their most powerful instruments.
The basic conceptual difference between the closed economy and the open economy case is
therefore that in the latter, as a general rule, it cannot be assumed that there is one global
social welfare function to maximize for all countries. Even if theoretical arguments on how
the full application of laissez-faire principles might benefit all countries alike are common, it
is usually correct to assume more realistically that each country strives to maximize its own
social welfare function.
In the open economy context, as a first approximation, it is often assumed that a small
country assumption holds, according to which trade policy measures taken in a country not
endowed with relevant market power are not expected to have any impact on the world price
of a given product. If a country’s Government decides to subsidize an industry, it will try to
do so by maximizing national producers’ benefits while minimizing the costs to domestic
consumers. A clear textbook exposition of mainstream theory on subsidies in the case of an
open economy, in the framework of a partial-equilibrium, comparative statics model, can be
found in Gandolfo (1998): its basic elements are illustrated in Box 1.2.
Yet cases in which each country’s actions do have an impact on international prices, and
more generally on other countries’ trade policy options and possibly on the very state of
international trade relations, are quite common in the real world. They are also more
interesting from the vantage point of an analysis aimed at evaluating different strategic
6
development policy options in the era of globalization. Moreover, in international and
interstate relations, non-economic (strictu sensu) arguments often weigh heavily in policy
makers’ welfare functions. Among them, geopolitical, military, ideological, religious and
prestige considerations might be as strong as or stronger than purely economic and social
considerations. In the international arena, the relevance of non-economic arguments in the
policy makers’ welfare function is much higher than in the case of a closed economy. As a
result, the chances that a conflictive, interventionist strategic-trade approach will be preferred
to a cooperative, mutually Pareto-efficient approach are also quite real.
Box 1.2
Definitions of subsidies in economic theory
From both a formal and a functional point of view, subsidies and taxes are opposite and equivalent:
subsidies can be considered as negative taxes, and vice versa. According to Samuelson and Nordhaus (2001, p.
78): “If taxes are used to discourage consumption of a commodity, subsidies are used to encourage
production…The general rules for subsidies are exactly parallel to those for taxes”.
In the Palgrave Dictionary of Economics, under “tax and subsidies” the focus is on corrective or
Pigovian taxes and subsidies “that could in theory be used to bring marginal private costs or benefits more
closely into alignment with marginal social ones” (Eatwell et al. 1989, p. 608). The need for such an alignment
is due to the presence of externalities (economies or diseconomies) operating at the margin. Externalities may be
between enterprises, caused by technological interdependence and other inter-firm linkages that may be
conducive to the realization of economies of scale and scope, or between consumers, where consumer A’s utility
function contains a variable also present in consumer A’s utility function. Externalities can also be between
firms and consumers, and a number of complex combinations are possible.
According to theory, consumers and firms’ optimizing behaviour drives towards the equalization of
both private costs and benefits. If the latter do not coincide with social costs and benefits, a market failure arises,
i.e. an overall Pareto-efficient equilibrium is not reached.
In the abstract world of traditional economic theory, where each agent can endlessly and at no cost
negotiate with every other agent, the most efficient taxes would not go to the state, but would be equivalent to
subsidies paid by one agent to the other in such a way that both end up in a Pareto-efficient equilibrium position.
In practice, due inter alia to the presence of significant asymmetries of information and transaction costs,
corrective taxes and subsidies enacted by the Government under its unique authority are the only solution. This
is true even if such taxes/subsidies are to be seen as second best solutions to be weighed against other second
best solutions such as licenses, quantitative restrictions and outright prohibition. It is important to point out that,
according to the above argument, corrective taxes/subsidies are always about achieving efficiency in a Paretian
sense, not equity. Therefore the issue of distributional equity should in principle be tackled by means of an
entirely different category of taxes/subsidies, along with other policy tools.
1.3.
Operational taxonomy
In theory, welfare is maximized when prices equal social marginal costs. If the market does
not spontaneously reach this equilibrium, Governments should intervene with taxes/subsidies
to correct market failures. In practice, marginal costs, and a fortiori social marginal costs, are
hard to evaluate, and the benchmark to define subsidies is provided by some kind of reference
prices, usually identified with opportunity costs. In the case of tradables, world market prices
are obvious candidates. However, interventionist policies in producer countries holding a
sizeable share of world markets do have an impact on world prices, and thus the latter cannot
simply be considered exogenous. Domestic prices of nontradables are even more obviously
affected by subsidy policies. Moreover, the reciprocal interaction of subsidy programmes and
other proactive policies at the domestic and the international levels further complicates the
task. In sum, the identification and evaluation of reference prices to define and quantify
subsidies is often not a straightforward exercise.
7
Taking this caveat into account, a pragmatic operational definition of subsidies is as follows:
“Subsidies comprise all measures that keep prices for consumers below the market level or
keep prices for producers above the market level or that reduce costs for consumers and
producers by giving direct or indirect support” (de Moor 1997, sub-section 2.3). On this
basis, an operational taxonomy based on the method of enacting the subsidies and consistent
with the WTO Agreement on Subsidies and Countervailing Measures, can cover four types of
subsidies (ibid.):
1.
Budgetary subsidies, which can be direct (grants or payments to consumers or
producers) or work through the effect of preferential tax policies (such as tax credits,
exemptions, allowances, exclusions, deductions, and rate relief, among others);
2.
Public provision of goods and services below cost, such as the provision of
infrastructure and complementary services and government R&D expenditures);
3.
Capital costs subsidies (preferential loans, loan or liability guarantees and debt
forgiveness);
4.
Policies that create transfers through the market mechanism, be they domesticoriented (such as price regulation, quantity controls, government procurement, and ad
hoc legislation) or trade-oriented (such as export and other types of subsidies);3
Some of these categories of subsidies are more difficult to evaluate than others. Budgetary
effects of tax policies can be measured with the revenue-foregone method (ex-post), the
revenue-gain method (ex ante), or the outlay-equivalent method, which estimates the cost of
providing the same benefit through direct spending. In practice, the revenue-foregone method
is the most widely used, but it is clear that the three procedures are not fully equivalent and
might lead to different estimates, thus suggesting that international comparisons should be
taken with a grain of salt. Public provision of goods and services can also raise difficult
computational and even theoretical difficulties. For instance, if the marginal cost of
infrastructure is virtually nil, it might be argued that the optimal option is in fact to charge a
zero price, and that such a pricing policy does not in fact amount to subsidization in the
proper sense. In practice, however, it is usually sounder to charge users at least part of the
recovery costs for services to shape their behaviour in order to internalize these costs to some
extent. Allowing, often not explicitly, public companies to generate a lower than market rate
of return is also a form of subsidy, but identifying the proper reference price (in this case, the
proper interest rate) is not an obvious exercise. Domestic market-mediated subsidy policies
encompass such a variety of price- and quantity-based measures that getting the overall
picture can be a complicated task. Finally, trade-oriented subsidies are a particularly
important category, because they directly affect international competitiveness and are the
favourite tools of neo-protectionist policies (section 2).
3
de Moor 1997 also lists in this category all tariff and non-tariff barriers. However, not all trade-oriented
policies represent subsidies in the operational sense (even if, from the point of view of domestic producers, they
are in a way equivalent).
8
1.4.
Estimating the magnitude and incidence of subsidies
Various methodologies have been proposed to measure the magnitude of subsidies. The most
important include:
- Nominal rate of protection (percentage difference between the domestic and the world
price);
- Effective rate of protection (percentage difference between value added per unit of
output at domestic and world prices);4
- Nominal rate of assistance (percentage difference between unit gross returns to
producers for domestic output and the world price);
- Effective rate of assistance (ERA); and
- Producer/consumer subsidy equivalent (PSE/CSE).
The last two indicators are particularly apt for international comparisons. The ERA is the
percentage difference between value added per unit of output at domestic and world prices,
taking into account all subsidies, taxes and other government actions that affect the price of a
product even indirectly. Its inclusiveness makes the ERA a very complete indicator, but all
the data required to calculate it are often not available.
The PSE, developed by the OECD, is more operational. It measures the value of transfers to
domestic producers at current levels of production, consumption and trade, using world prices
as benchmarks. It includes market price support, direct payments, reductions in input costs,
general services and other indirect support, but excludes some specific subsidies such as
payments to offset social security costs and grants to cover subsidence damage. The PSE
concept has been used by van Beers and de Moor (2001) in their overall evaluation of
subsidies worldwide, the main results of which are reported and discussed in section 3. It was
also adopted by the WTO to monitor the implementation of the Uruguay Round Agreement.
The WTO applies a slightly different procedure, calculating an aggregate measure of support
(AMS) on the basis of a fixed reference price (in order to measure the progress towards
liberalization over time). A crucial detail, moreover, is that the AMS, contrary to the PSE
concept, does not include R&D subsidies.
Finally, it is worth briefly mentioning the problem of how to evaluate the direct and indirect
incidence of subsidies. Clearly, if measuring the sheer magnitude of subsidies is not easy,
assessing their full impact on the domestic and the international economy is far more
difficult. Ideally, a complex computable general equilibrium (CGE) model would be required,
but this technique is not usually feasible. Usually, in policy-oriented, practical exercises, only
a limited number of economic, social and environmental indicators are considered,
sometimes in a multi-criteria framework. Among the most commonly used economic
indicators are GDP or productivity growth rates (either relative to the whole economy or to
specific sectors), domestic or international market shares, and employment data. Statistics on
emission reductions and on the depletion of natural resources can be used as environmental
indicators, and the Gini coefficient and the incomes of specific groups of population as social
indicators. Even when applying this less ambitious approach, however, analysts should strive
to gauge heuristically the most evident and relevant intersectoral linkages and other indirect
effects of subsidies.
4
The effective rate of protection is a concept belonging to economic theory that applies essentially to tariffs. It
has only an indirect relation with concrete policy measures such as subsidies.
9
2.
THE POLITICAL ECONOMY OF SUBSIDIZING SERVICES
2.1.
Sub-optimal subsidy policies as a social choice problem
If subsidies, like all interventionist attempts to tinker with the free working of market forces,
cause various types of distortions, constitute under the most optimistic assumptions a second
best solution, and are bound to disrupt the international trade scenario, thus possibly
thwarting the overall process of globalization, with all its universal benefits, why then are
they so popular and hard to eliminate? This is the traditional, rhetoric question that theoretical
economists ask now and again. The answer, usually, is rooted in ignorance. The general
public ignores the economic effects of subsidies and sees only their superficial, immediate
benefits. More specifically, most people (consumers/taxpayers) are left in the dark on the
ultimate, negative consequences of interventionist policies. Thus, they tolerate politicians
imposing subsidies that only benefit relatively small interest groups at the expense of the
majority of the population. Worse still, this is not even a zero-sum game, where some
people’s losses are exactly compensated by other people’s gains, as it leads to a Paretoinefficient equilibrium where everybody is, on average, worse off than if the market had been
left free to operate spontaneously. Such an unsatisfactory state of affairs, a fortiori in a
democracy, is only possible because politicians are in fact working not for the common good,
but on behalf of special interest groups or lobbies. These groups are small, homogeneous,
organized, and informed enough to realize the particular benefits they reap from subsidies,
and they are also in the position to control enough votes to make or kill any politician’s
career.
Moreover, both beneficiaries and institutions get easily “addicted” to subsidies (van Beers
and de Moor, 2001, chapter 4). It is not only a matter of psychology. Once a sector is
subsidized, relative prices are altered, and producers rationally invest in fixed capital to a
larger extent than would have been the case in a no-subsidy situation. If subsidies were
phased out, these investments would become uneconomic and could lead to a wave of
bankruptcies. In other words, the subsidy policy pushes producers to engage in rent-seeking
behaviour, and a policy change would impose significant adjustment costs on them.
Institutions, such as government bodies, are resistant to change – actually, their very
existence is justified by some theorists as a means to reduce and make more manageable the
instability, uncertainty and unpredictability that would prevail if all economic interactions
were left to spontaneous market forces, due among other things to asymmetric information
and transaction cost problems. Institutions are also path-dependent, i.e. once a bureaucratic
function and a corresponding body have been created, their elimination would not be socially
and economically costless (think, for instance, of the implications of removing agricultural
subsidies programmes, which presently employ a numerous and specialized public
workforce).5
5
The concept of path-dependency in institutions is analogous to the one familiar to technological economics
(see van Beers and de Moor 2001, para 4.4.)
10
Box 2.1
International trade liberalization, national states, and subsidies:
A political economy approach
O’Brien (1997) has proposed a political economy and institutional approach that stresses the tensions
between multilateral trade liberalization and the prerogatives of national States on the issue of subsidies.
According to O’Brien: “The ability to tax and spend is fundamental in our understanding of the modern state”
(p.17). Control over taxation was at the core of both French and American revolutions. A true State in the
traditional sense must be able to collect revenues and spend them without recourse to outside powers. Broadly
defined as whatever form of public expenditure that supports the production of goods and services, expenditures
can include “just about every function that governments undertakes from running a police station to providing a
tax break for corporations” (p.18). Yet nowadays the concept of state sovereignty is undergoing important
changes, with the unprecedented relevance and powers of the WTO and other super-national and/or multilateral
trade bodies.
International attempts to regulate and curb subsidies challenge the traditional understanding of the very
concept of state, as they restrict its ability to dispense revenues, according to four general trends which inform
the overall drive towards (neoliberal) globalization:
1. Liberalization of the state, which reduces Governments’ involvement in the economy. A common
and still largely WTO-compatible response on the part of national States is to reduce direct assistance to goodsproducing firms, while concentrating on broad infrastructural expenditures on social and other services that are
indirectly instrumental in shaping the overall socially determined competitiveness of each national economy as a
whole (among them social support, education, research, welfare, transport and communication). This neoprotectionist trend, however, is tempered by the reluctance of OECD countries to phase out direct support
programmes for key sectors such as agriculture, audiovisuals or aerospace (section 4).
2. Liberalization versus autonomy. In many countries, liberalization and globalization trends produce a
major contradiction. Social and economic groups that are failing or perceived as failing push in the domestic
political arena for a more aggressive posture vis-à-vis other countries, accusing them of unfair trading practices
and pressuring for a combination of protection at home and attack abroad. As by now the goal of tariff reduction
has been largely achieved, trade rivalries and disputes are deviated, with augmented emphasis on other direct
and indirect trade policy instruments: “Subsidies are a particularly sensitive area because most of the
controversial ones have been established to pursue domestic goals rather than achieve competitive advantage…
(therefore) the attack on another government’s domestic policies is an attack on its policy autonomy” (p.19).
3. Transfer of decision-making authority. This authority appears to be progressively slipping away
from States’ grasp and shifting toward regional (i.e. the EU) and multilateral (i.e. the WTO) bodies. It is only
apparently paradoxical that the United States, the country which has pushed most for the creation and
progressive empowerment of WTO and more generally for the neoliberal globalization model, is in fact the one
that is the most reluctant to submit to this transfer of decision-making. This evident contradiction, taking into
account the weigh of the United States, creates an intrinsic and ever-resurfacing tension in the globalization
drive.
4. Erosion of democracy and accountability. The above-mentioned tensions are exacerbated by the fact
that multilateral trade bodies, bureaucracies and courts are not elected bodies, even if they are not intrinsically
undemocratic either. Moreover, once agreed upon, treaties are not regularly subject to popular review, and in a
way they resemble a long-term constitution more than legislation subject to review. Physically, they are located
far away from most of the world populations and tend to operate on highly technical grounds. Thus, they are
seen by many as the domain of obscure technocracies, on which it is extremely difficult to exercise any form of
democratic control. Yet, they are increasingly bestowed with decision-making power, even on locally sensitive
matters such as the amount to be spent to support rural communities. Thus, attempts “to create institutions
which could depoliticize subsidy issues and turn decisions about tax expenditure into technical exercises” (p.20)
also create contradictions, as well as rational and irrational popular fears.
The explanation outlined above contains more than a grain of truth. However, it reflects a
universal problem faced by social choice in general, and should not produce a naive
interpretation of social and political mechanisms leading to the implementation of subsidy
programmes. There are a number of features in social life that lead to sub-optimal equilibria
11
from the point of view of common welfare and that could ideally be superseded in favour of
Pareto-efficient solutions if every member of society behaved in a perfectly socially virtuous
fashion. For instance, public expenditure on police forces is a social cost that could easily be
eliminated if no citizen indulged in criminal activities. It can also be seen as an inefficient
policy tool enacted by the state to protect the interests of one part of the society (noncriminals) at the expenses of another part (criminals). Many other socially useful, albeit
inefficient subsidies, can be seen in a similar framework.
In sum, subsidies, like virtually all state policies, as well as the actions of other social and
economic agents, are not, and cannot be expected to be, aimed at maximizing the ultimate
welfare of humankind as a whole. In a less than perfect world, they must rather be seen and
evaluated realistically as tools to boost the welfare of some social groups. Even taking such a
pragmatic approach for granted, the sheer magnitude of subsidies, and the scale of the social
and environmental costs they often impose on both developed and developing countries due
to their perverse effects, cannot be overstated (section 3). Therefore, the existence and
persistence of socially counter-productive subsidies should not be taken as an inevitable fact
of life, and a broad, if gradual, reform of overall subsidy policies should figure prominently
in policy makers’ agendas.
2.2.
Cooperation and conflict in international trade policy rivalries
As mentioned in the preceding subsection, economic theory contemplates the possibility of
reaching a solution based on cooperation, leading to a mutually preferable equilibrium with
respect to the outcome that would arise from the adoption of conflictive strategic trade
policies.6 Still, subsidies and other instruments seen as trade policy tools should not, as a
general rule, be seen in a universal Pareto efficiency framework, but in a rivalry-based
strategic trade policy framework. It might be true to state that reciprocal and universal
elimination of state-induced trade distortions (worldwide trade and financial liberalization)
would improve universal welfare, but this is not by itself more meaningful than stating that
universal and full disarmament would do the same. In both cases two fundamental problems
arise related to the feasibility of bribing and to the concept of trust respectively. The first
problems stems from the fact that for a (by assumption) globally Pareto-efficient
liberalization measure to be undertaken bilaterally or multilaterally, every State must gain
something, or at least it must not lose. It is often very difficult in practice to compensate the
losers, even where theoretical global gains from liberalization are huge. The problem of trust
is self-evident and relates, among other things, to issues such as the imperfect and
asymmetric information on rivals’ intentions and, possibly, hidden agendas.
In the international arena, there is always the possibility that a trade problem will be solved
not through meaningful negotiation but through de facto unilateral imposition of the will of
the stronger side by means of war, military threat or milder trade-related or broader economic
and political pressure. Even in these cases, there is the theoretical possibility that the result
could be a welfare improvement in both countries (i.e. the case of a strong trade partner
forcing a weaker one to phase out inefficient protectionist measures that were effectively
making that country worse off than an alternative scenario of no trade policy intervention
whatsoever). However, it is fair to assume that, especially in the unilateral imposition case,
the stronger side only maximizes a national rather than a global welfare function. Thus, an
6
Actually, the main justification for the existence of quasi-universal multilateral trade agreements is precisely
that of limiting the damage caused by this form of “tragedy of the commons”.
12
outcome with an improved welfare level for the weaker side could only arise as a casual byproduct of the stronger side’s unilateral and selfish action, and only if the level of selfdefeating inefficiency of the weaker side’s previous trade policy was extremely high. Such an
outcome, albeit not impossible, should in fact be considered an oddity.
Box 2.2
Internalization, subsidies and strategic trade policies
According to Busch (1999), a relatively simple game theory framework helps to understand the basic
features of trade-strategic rivalry and its possible outcomes in the present WTO-conditioned international trade
scenario. Each potential subsidizer knows that engaging in an aggressive interventionist policy in support of
“national champions” can provoke a retaliation on the part of rivals, with the risk of unleashing a trade war,
which might end up leaving both parties worse off than if neither had interfered in the market. Thus, seeking to
maximize national welfare, States weigh the expected benefits from subsidies against expected costs. The latter,
of course, can never be nil, but the main issue is the likelihood that they might skyrocket due to the reaction of
trade rivals. Benefits are essentially related to two categories of externalities: linkage externalities, stemming
from downstream and upstream scale economies, and spillover externalities, stemming from the diffusion of
technology among related industries. It is crucial to gauge not only the magnitude of these externalities and the
ability of national firms to realize them, but also to what extent they can be captured by national agents, without
spreading outside the national borders, i.e. to what extent they can be internalized. If internalization prospects in
a given industry are poor (usually because its structural features are such that technology advances originating in
one country rapidly spill over borders and are easily and cheaply appropriated by other countries), supporting
the industry would equate to subsidizing trade rivals, who would act as technological free riders.
According to Busch, policy makers are less irrational than economists sometimes portray them to be
and do not always engage in a subsidy frenzy. On the contrary, they tend to intervene in an industry (if they
intervene at all) only to an extent proportionate to their “rational expectations” on key variables, such as the
magnitude of the externalities, their “capturability” on the part of national industry, and the degree of
internalization of expected benefits. In a two-country strategic trade game between rivals of comparable
strength, this attitude can typically lead to various different outcomes.
If only one State can capture and internalize the externalities while the other cannot, the first one will
heavily subsidize the industry until foreign competitors are swept away from the market. This outcome is all the
more likely because, as the prospects for its own industry are not favourable, the other State will probably
choose not to retaliate. Such a dichotomized incentive structure, leading to drastic subsidies shelling, is however
less frequent than four other configurations, which lead to various forms of limited intervention.
If externalities are high and both countries are in a position to capture and internalize them, rivals are
driven to heavily subsidize their respective industries. This policy stance, however, is bound to provoke an
escalation, leading to diminishing returns for both players and to the risk of a full-fledged trade war. Under this
threat, both States may seek to cooperate, agreeing on lowering the level of subsidies. The US-Europe civil
aircraft rivalry can be interpreted along these lines.
If both countries can capture the externalities but cannot internalize them, they will initially try to free
ride and neither will subsidize the industry. As a result, the industry will develop too slowly, hurting both
countries. Contrary to the previous case, therefore, they might end up brokering a solution based on both
countries increasing subsidies, creating a different form of limited intervention equilibrium. This has been the
case in the US-Japan semiconductors rivalry.
If one country can capture, but not internalize the externalities, and the other cannot capture them, the
first player will subsidize the industry, but only to a point, as the country is aware that part of the benefits will
inevitably spill out of its borders and in the future might even help the rival to exploit a “late mover” advantage,
making it a formidable competitor. An example of this configuration is the US-Japan High Definition TV
rivalry.
Finally, if neither country can capture or internalize the externalities, there is no incentive to subsidize,
and the rational strategic choice is simply non-intervention.
Apart from this extreme case, individual States and coalitions of States are usually
interlocked in complex strategic trade games that alternate and combine moments of intense
rivalry with moments of cooperation. An interesting approach to the analysis of subsidies to
13
R&D and other activities, in the framework of a two-country game for supremacy in high
technology industrial sectors, is presented in Box 2.2.
2.3.
Subsidies to services as strategic trade policy tools and the crucial role of
education and R&D
Subsidies to service activities, more so than subsidies to the primary or secondary sectors, can
play a very important role in any strategy aimed at accelerating growth and enhancing the
international competitiveness of a country’s entire economic system.
Not only has the relative weight of services been expanding to cover an increasing share of
total economic activity both in developed countries and in many developing countries,
material production activities themselves (many of which are tradable) are increasingly
dependent on activities classified as services (many of which are still non-tradable). The
availability, efficiency and cost of infrastructural and other key services are also increasingly
important in determining the international competitiveness of the agricultural and industrial
sectors and thus of the national economy as a whole. Accordingly, subsidizing certain key
services that represent inputs and/or the availability of which constitute key enabling
conditions for commodity production is equivalent to indirectly subsidizing the latter, with
obvious trade policy implications. This argument applies not only to subsidies earmarked for
traditional services sectors such as energy and transport, or to more modern infrastructural
services such as telecommunications, but also to basic services such as education and health.
Subsidies to education and health services have a direct impact on human capital formation.
While in advanced countries the impact of subsidies on international competitiveness in
technologically advanced sectors works mainly via the availability and cost of skilled labour,
in poor countries subsidies to basic services also affect the very working ability and, quite
directly, the material productivity of unskilled labour (think, for instance, of the devastating
impact of AIDS on African rural productivity and of how it might be mitigated by an
improvement in public health services).
Still, as the role of knowledge in the modern economy cannot be overemphasized,7 higher
education and R&D services have a unique, crucial and universal impact on each country’s
productivity and competitiveness. Even if the relationship between cause and effect is
complex and non-linear, it is safe to state that the set of activities aimed at increasing and
spreading a society’s stock of knowledge (including, for instance, education), and R&D
activities in particular, are the main source of innovation and productivity growth, and
therefore ultimately of economic development at large.8 Besides their key role as productivity
boosters, education and research also share two other important characteristics; both are
plagued by externalities, which mean there is bound to be underinvestment by private agents
7
Investment in R&D is the centerpiece of “new growth theory”. See, for instance, Helpman and Grossman
1991, Aghion and Hovitt 1998.
8
Several empirical microeconomic studies have robustly confirmed the linkage between R&D and productivity
growth at the firm and industry level (Griliches, 1998). A more controversial question is the following: to what
extent are subsidies to R&D additional with respect to private R&D (i.e. end up producing more R&D, as
intended by policy-makers), and to what extent, on the contrary, do they “crowd out” or substitute for private
R&D, thus failing to reach their ultimate goal? Many recent studies have explored the “additionality effect” of
R&D subsidies, both theoretically and empirically, reaching different conclusions on its magnitude under
different conditions and policy scenarios (see, for instance, Guellec and van Pottelsberghe dela Poterie 1997,
Busom 2000, David, Bronwyn, and Toole 2000, Lach and Sauer 2001). However, there is little doubt that, by
and large, R&D subsidies do boost overall R&D at the national level (and thus, ultimately, economic growth as
well) to some extent.
14
left to the vagaries of the market, and both are primary candidates for state support.
Moreover, the negative effect and perverse distortions commonly associated with subsidies
are likely to be minimal, if any, in the case of education and R&D.
The case in favour of subsidies for education is usually, and correctly, put forward mainly on
social grounds. Yet public support for education can also be seen as a tool to boost a
country’s international competitiveness. Actually, Governments, besides subsidizing part of
the private sector’s R&D effort, also carry out their own R&D in public research institutes
and finance higher education. As a result, each country’s enterprises system, seen as a whole,
reaps two categories of competitive benefits from public subsidization of R&D. The first
category is constituted by the direct cost saving stemming from the subsidization of
enterprises’ own R&D. The second category of benefits is indirect, and stems from the
overall educational and technological spillovers of the R&D carried out by public bodies and
by Governments’ funding of higher education. By this token, the huge gap between per capita
public subsidies for education in developed and developing countries amounts to an unfair
competitive disadvantage for the latter.9
Box 2.3
The distorting impact of subsidies to developed countries’ construction firms
Competition among large and small firms in construction services is intense, yet a high level of public
intervention characterizes the sector. In developing countries, in particular, dualism is the norm: on one hand,
small-scale private construction works are carried out by local firms in an anarchic and unregulated fashion; on
the other hand, the big business of large construction projects is dominated by state procurement, complex
regulations and public financing. Especially in poorer countries, the role of bilateral and multilateral donors is
also paramount.
Developing countries’ firms, due to a number of structural weaknesses, are at a severe disadvantage in
competing for large construction works in their own countries, let alone exporting their services abroad,
although some interesting exceptions do exist. Under these circumstances, donors’ financing for large
infrastructural projects in developing countries is almost completely captured by a few large transnational
corporations from developed countries, through de facto quasi-monopolistic practices. Thus, the bulk of donor
financing for construction projects in developing countries amounts to an indirect form of subsidy for developed
countries’ own construction firms, additional to other specific subsidies they usually receive from their
Governments (for R&D, feasibility studies, market development and the like).
The severe difficulties encountered by developing countries’ construction firms in competing in their
own countries’ markets against foreign firms from developed countries, due among other things to the
pervasiveness of various forms of subsidies and to the existence of other unfair trade practices, were underlined
in an UNCTAD Expert Meeting held in Geneva in October 2000. In the outcome of the meeting, experts agreed
that developing countries had little alternative but to strengthen the competitiveness of their national
construction firms through strong interventionist policies, including various forms of subsidies: “The
development of a strong domestic construction sector (in developing countries) should be supported by
instruments of industrial policy… adequate funding for research and development is an essential component of
such policies…”(UNCTAD 2000a, p.2). Given the uneven subsidization potential between rich and poor
countries, the international community should substantiate the special and differential treatment principle by
allowing developing countries ample room to enact support policies, while curbing unfair trade practices on the
part of developed countries. The GATS negotiations, in particular, should “address measures which impede the
ability of developing country firms to compete in their own markets and in those of other countries, including
subsidies…Developing country Governments should identify…the main trade barriers affecting their trade in
construction services…These include restrictions on… subsidies and discriminatory tax policies…” (UNCTAD
2000a, p.4).
9
As a matter of fact, it could paradoxically be argued that public support for education in OECD countries
should be gradually phased out in the framework of global progressive liberalization.
15
Especially in the domain of high technology sectors, subsidies to services activities constitute
one of the most important tools of strategic trade policy. High-tech industrial sectors do
produce tradable goods, but their relationship with services is very close. The strategic
relevance of high-tech industries is not so much in their manufacturing and trading aspects,
but in their unique position as intensive users and creators of advanced technology, first-class
human capital and, more generally, knowledge. High-tech industry is the domain of
externalities par excellence, of which technological spillovers, backward and forward
linkages, economies of scale and scope, and path dependency are only the most obvious
examples. In a long-run, strategic perspective, for an industrialized or even a semiindustrialized country, the development of high-tech industrial sectors is the most decisive
determinant of the overall development of the national economy taken as a whole. As a
consequence, it is also a decisive component of the country’s international trade
competitiveness, and thus the privileged field of intervention for strategic trade policy
makers. It is therefore not surprising that “developed and developing states have been
jockeying for a foothold in high-technology industries, the spoils of which may be had by just
a few lucky winners. These commercial rivalries are fuelled by concern that gains and losses
may be path-dependent, and that success in these industries may confer an insurmountable
lead in building the ‘economy of tomorrow’” (Busch 2000, p.2).
Box 2.4
Subsidies to the health sector as an export promotion tool in developing countries
One of the main sectors subsidized in all countries, for obvious social reasons, is that of health services.
Nevertheless, international trade in health services is expanding, and it could offer new export opportunities to
developing countries. As a matter of fact, in such a human-capital-intensive sector, many developing countries
can effectively overcome their technological disadvantages, capitalizing on the availability of a cheap and
qualified workforce and achieving international competitiveness in a number of niche subsectors. Moreover,
potential synergies with other services, such as tourism, also offer promising development potential (UNCTAD
1997, p.10).
As in other sectors, these opportunities are conditional on the fulfilment of many requisites, some of
which can be resolved through multilateral trade negotiations (for instance, the liberalization of movement of
natural persons and an improvement in the domain of international recognition of qualifications and diplomas
and of the cross-border portability of health insurance). Still, strong national support for the health sector and
the elaboration of effective export strategies are also essential, as showed by the successful experience of some
developing countries (UNCTAD 1997, p.1). The practicability of effective proactive strategies, with their
important export-promoting component, requires on the part of developing countries the capacity to earmark
scarce resources for the health sector and a measure of restraint on the part of developed countries in subsidizing
(often indirectly) their own export-oriented private health firms. The last point was highlighted by experts
participating at the 1997 UNCTAD expert meeting on international trade in heath services: “(improved)
disciplines on subsidies could increase the participation of developing countries in global trade in
services…Experts expressed the view that issues such as subsidies, safeguards and government procurement
deserved more attention in the future…” (Chairman’s SUMMARY of the experts’ informal discussion, pp.5-10,
UNCTAD 1997).
High-tech industrial sectors are intensive users of highly specialized R&D and of other
knowledge-intensive services activities – which as a group can be dubbed as advanced hightech support services – and, to be efficient, they must rely on a complex and costly public and
private service infrastructure. Subsidies to these services constitute strategic trade-policy
tools ultimately aimed at enhancing the overall international competitiveness of each
country’s economy taken as a whole. The feasibility of such an aggressive technology and
trade policy approach, from the point of view of international trade law, can be conditional
upon the formal classification of R&D and other advanced high-tech support services. These
16
can be internalized in the high-tech firm: in this case, they are formally classified as ancillary
activities carried out in the framework of industrial material goods production, and therefore
they often fail to be considered statistically as services in their own right. Alternatively, as
happens ever more frequently, such service activities are externalized, and these crucial nonmaterial inputs are bought by the high-tech industrial firm through market transactions. In
this latter case, advanced services should be statistically classified as part of the broader
services sector. This classification difference is very important to specify the applicability of
multilateral trade liberalization rules, as in some cases policy measures such as subsidies
might be WTO-compatible if they are formally classified as belonging to the services sector
and WTO-incompatible if they belong to the industrial sector.
3.
SUBSIDIES IN DEVELOPED AND DEVELOPING COUNTRIES: A REVIEW
OF QUANTITATIVE EVIDENCE
3.1.
The difficult task of estimating global subsidies
Various factors (among them the multiplicity and pervasiveness of subsidies in virtually all
countries, developed and developing; the variety of direct and indirect instruments utilized by
Governments to implement – and often also to disguise – subsidy policies; and a relative
understatement of their importance as a major economic, social and political phenomenon)
help to explain why comprehensive attempts to estimate the overall magnitude of subsidies
worldwide are relatively rare and do not go very far back in time. In the 1970s and 1980s,
global awareness on environmental problems, and the emergence and popularity of the
concept of sustainable development, led to new interest in analysing the unintended and often
perverse environmental impact of economic policies, among them subsidy programmes. In a
1992 World Bank study, Shah and Larsen warned that subsidies to the energy sector of the
order of hundreds of billions of dollars worldwide were a major cause of pollution, let alone
economic inefficiencies. Estimated global subsidies to agriculture, energy, water and road
transport stood at US$ 700-900 billion, and they often encouraged ecologically harmful
behaviour and usually had a negative distributional impact (as they disproportionately
benefited relatively rich population groups). Another estimate by Myers, who applied
partially different criteria to include various forms of indirect subsidies, put the total of
perverse subsidies at US$ 1,400 billion. Two OECD (1997, 1998) studies confirmed that
large subsidies to agriculture, industry and transportation in developed countries routinely
cause pollution and excessive use of natural resources. The remainder of this section reports
briefly on and illustrates the estimates proposed in the most recent comprehensive study on
subsidies worldwide (van Beers and de Moor 2001).10
3.2.
More than a trillion dollars per year
Global public subsidies are more than a trillion dollars per year (US$ 1,065 billion), or four
per cent of world GDP. This amount corresponds roughly to the equivalent of the combined
GDP of all low-income countries, or to the total external debt of developing countries. OECD
10
In 1995, under the guidance of the former OECD Secretary-General, Emile van Lennep, the Earth Council
embarked on an initiative to evaluate the costs and impacts of public subsidies, with the main goal of unveiling
their negative environmental consequences. The initiative grew to a progressively more structured research
project, with the constitution of the Institute for Research on Public Expenditure, and to the publication of
various reports. Van Beers and de Moor have been contributing to the Earth Council initiative for a number of
years. In their most recent book, along with a thorough analysis of the political economy of subsidies and
proposals for subsidy reforms, they synthesize the main quantitative results of their research on subsidies, many
of which had been previously published in van Beers and de Moor 1999 and in OECD 2000.
17
countries spend more than twice as much on subsidies in relation to developing countries
(US$ 725 billion against US$ 340 billion), but much less in proportion to GDP (3.4 per cent
and 6 per cent respectively) (van Beers and de Moor 2001, subsection 3.3). These figures
inspire two basic remarks.
First, absolute and, a fortiori, per capita expenditure on subsidies is much higher in
developed than in developing countries. It does not seem to matter how inefficiently this
expenditure contributes directly and indirectly to increasing the international competitiveness
of national firms (see the arguments developed in section 2 on strategic trade policies). As the
financial capability to enact subsidy policies is unequally distributed worldwide, the political
will to pursue these policies on the part of developed countries contributes powerfully to
invalidating the abstraction of an even playing field in global trade and to perpetuating and
magnifying poverty and underdevelopment in developing countries.
Second, and only in part as a reaction to developed countries’ subsidy policies, the weight of
subsidies in developing countries is particularly high. Many of these subsidy policies
(although by no means all of them, as some selective subsidy policies are necessary tools of
development strategies) are likely to contribute to some extent to inefficiencies and
distortions in developing countries’ economies.
Subsidies are widespread in all sectors, but half of total subsidies are earmarked for natural
resources sectors. Such a large share is particularly perverse and paradoxical, as it encourages
the exhaustion of irreplaceable and non-renewable natural resources which, according to any
official political discourse, should be preserved and defended as an utmost priority. This
contradiction is particularly strident in the developing countries, where subsidizing the
overexploitation of natural resources in a short-sighted attempt to boost production and
exports and often leads to the classical fallacy of composition outcome and ultimately to the
additional boomerang effect of reducing their price and aggravating the country’s terms of
trade.
Among natural resources sectors, agriculture gets the lion’s share (US$ 400 billion
worldwide). OECD countries, in particular, channel half of all their subsidies to agriculture.
From the point of view of the focus of this paper, it is interesting to note that, apart from
agriculture, the three most subsidized sectors all involve services: water, energy and road
transport. These three services sectors absorb almost half of total subsidies worldwide and an
even higher share (mainly accounted for by energy) in non-OECD countries (table 3.1) Water
collection and distribution is subsidized more heavily in developing countries (US$ 45 billion
per year) than in OECD countries (US$ 15 billion).11 Energy is also heavily subsidized (US$
240 billion a year). Road transport is another highly subsidized sector, absorbing more than
four times the amount of subsidies of the manufacturing industry. With respect to the latter, it
is interesting to note that the quasi-totality of manufacturing subsidies (US$ 55 billion)
benefit industries in OECD countries, thus enhancing their global competitiveness and
jeopardizing that of developing countries.
11
A significant part of water subsidies is earmarked for irrigation. These water subsidies overlap to some extent
with subsidies to agriculture (see subsection 3.3).
18
Table 3.1
Percentage share of three large services sectors in total public subsidies per
year
(US$ billion)
OECD Non-OECD World
Energy, water, road transport
41
68
49
Agriculture
46
19
38
Others (including manufacturing industry)
13
13
13
Note. Estimates refer to the 1994-1998 period.
Source: van Beers and de Moor 2001, table 3.1.
3.3.
Subsidies to water distribution services
Global water reserves are not unlimited and are unevenly distributed. Thus, water is actually
scarce in many regions (namely the Middle East and North Africa) and subregions. On
average, 70 per cent of water use is accounted for by agriculture, a figure reaching 90 per
cent in the case of low-income countries. Conversely, in high-income countries, almost half
of total water use is accounted for by industry. Global water demand doubles every 20 years,
spurred mainly by water demand from agriculture and industry. Access to safe water in poor
countries is inadequate, due among other things to losses and inefficiencies in irrigation and
distribution systems. Research has shown that price elasticities for water users are significant
and can be over one for non-residential (agricultural and industrial) users.
Due to the obvious economic, social, political and geo-political relevance of water,
Governments are heavily involved in water issues. Yet government interventions in water
distribution, and subsidies in particular, are often inefficient and/or counterproductive. On a
global scale, the main distortion they provoke is the under-pricing of a resource that is in fact
scarce, favouring waste and suboptimal uses. Water subsidies take different forms. Among
the most widely used are official prices set so low that they do not even cover marginal costs
for irrigation and recurrent (let alone capital) costs for drinking water; VAT exemptions; tax
exemptions; and soft loans for intensive irrigated agriculture.
In OECD countries, there are two main forms of water subsidies: grants and low-interest
loans. The weight of water subsidies is uneven and particularly high in Australia, Japan,
Turkey, and the United States. Most subsidies are earmarked for capital expenditures, with
grants covering on average between 20 and 40 per cent of total costs but reaching up to 80
per cent for irrigation projects. Many irrigation projects are not intrinsically economically
viable per se, and are carried out only because of the enormous subsidies involved. Several
studies have shown that irrigation tends to be extremely underpriced in OECD countries. On
the basis of an average subsidy rate of US$ 54 per acre, total annual subsidies for irrigation
for the 40-50 million acres of irrigated land in the United States can be estimated at US$ 22.5 billion (Repetto 1986. Moore and Mc Gucking 1988. De Moor 1997. Van Beers and de
Moor 2001), resulting in vast rents benefiting mostly large farmers. Moreover, water
subsidies encourage farmers to overproduce and thus obtain further government subsidies
under acreage reduction programmes. Water subsidies in the United Kingdom are in the order
19
of US$ 1.2 per year (Maddison et al. 1997). Total water subsidies in OECD countries are at
least US$ 15 billion per year (van Beers and de Moor 2001).
Water subsidies in the developing world are even higher, around US$ 45 billion per year, half
of which are for irrigation (De Moor 1997, Myers 1998). The two main justifications for
subsidizing water in relatively poor countries (to increase agricultural production and thus
food security; to help the poor to get access to clean drinking water) are virtuous in principle.
However, more often than not, subsidies fail to reach the poorest segments of the urban and
rural population, encourage inefficiencies and waste, and aggravate pollution. The fiscal
burden of delivering ever-scarcer and underpriced water to households, especially in large,
fast-growing and polluted cities, is very high and always on the rise. Part of the cost is due to
inefficiencies and losses caused by underpricing. Irrigation is also subsidized, with cost
recovery between 10 and 20 per cent on average in the 1980s (Repetto 1986), a figure
insufficient to cover even operating and maintenance costs. Subsidies to irrigation in
developing countries are estimated at about US$ 20 billion per year. According to studies by
the World Bank (1993, 1994) and others, the reduction and rationalization of water subsidies,
especially with respect to irrigation, would provide developing countries with economic,
fiscal, social12 and environmental benefits. On a global scale, however, it is economically and
socially difficult for developing countries to cut support for their own agricultural sector
(where a large part of the population still makes its living) by removing irrigation subsidies,
while developed countries – which would face much smaller social and political problems,
due to the tiny fraction of their population still living on the land – continue subsidizing their
own agricultural sector heavily through an array of market-distorting policy instruments, of
which irrigation subsidies are but one of many.
3.4.
Subsidies to energy services
Over the past 50 years, global energy consumption has quadrupled. The main energy
consuming sectors are electricity, industry, transport and households. Households account for
about 20 per cent of total energy consumption. Per capita use of commercial energy in highincome countries is three to four times the world average. The main energy source in
industrial countries is oil, followed by coal, natural gas and nuclear energy. Renewable
energy sources play a marginal role. Income and price elasticities of energy consumption are
quite significant.13
Experts’ estimates on the future supply of non-renewable energy sources diverge. The
mainstream opinion is that a sudden, catastrophic scarcity is an unlikely scenario, but existing
subsidies should be eliminated in order to let the market put a proper price on these resources
(quite apart from the crucial issue of adequately taking into account the cost of environmental
degradation caused by present energy use patterns).
The main environmental problems caused by energy production and use are the following:
-
Global warming;
Land and sea pollution caused by acid rain, dust and soot, and incomplete
combustion;
12
According to the World Bank, contrary to what many believe, water subsidies actually disproportionally
benefit the rich. Hence, they are distributionally regressive.
13
Income elasticity tends to be higher in developing than in developed countries, while the opposite is true for
price elasticity. If the overall economic linkages of energy use were considered, these elasticity estimates would
probably have to be revised upwards.
20
-
Loss of biodiversity;
Deterioration of water quality;
Health problems, mainly respiratory diseases;
Specific energy sector diseases (i.e. miners’ black lung);
Safety risks (mainly associated with nuclear energy and waste);
Land use conflicts;
Dust and soot problems.
Governments aggravate the problems by subsidizing energy production and consumption,
and often create biases favouring the most polluting energy sources. OECD countries tend to
subsidize energy production, while developing countries concentrate on subsidizing energy
consumption (OECD 1997, van Beers and de Moor 2001). Yet from a domestic
macroeconomic point of view, most oil and gas exporters effectively tax the energy sector to
subsidize the rest of the economy. Thus, it can be stated that the energy sector in most
countries is at the centre of a complex web of subsidy flows,14 usually regulated by heavy
state intervention, which sometimes lead to inefficient or even perverse outcomes.15
To subsidize energy services, Governments use various policy tools – such as grants,
mandatory regulations, training assistance, price controls and guaranteed markets – targeting
either energy producers or consumers. Relevant examples in the industrialized world are as
follows:
-
Direct grants to the coal sector in the United Kingdom, Germany, Japan, and Russia;
Tax subsidies (depletion allowances) in the United States;
Tax exemptions for consumers in the United Kingdom and Italy; and
Price support to domestic producers in Germany.
Various other forms of subsidized energy are also widespread, among them loan guarantees
and soft loans to energy producers, rules designed to allow public energy producer companies
to earn a lower-than-market rate of return, and below-cost public provision of key services
(i.e. R&D) to the energy industry.16
To sum up, OECD countries massively subsidize energy production and distribution. Energy
subsidies falling under the PSE sub-section 1.4 definition are estimated to be between US$ 18
and US$ 32 billion in the US,17 about US$ 10 billion in Europe, and US$ 6 billion in other
IEA countries, for a gross OECD total of between US$ 53 and US$ 67 billion.
Adding non-PSE coal and other subsidies (such as deficit payments to miners’ pension
funds), total OECD energy subsidies might run to US$ 70-89 billion. This figure might in
14
This kind of intersectoral cross-subsidization falls outside the definition and taxonomy of subsidies potentially
relevant from the point of view of multilateral trade negotiations (see section 1).
15
Increasing realization of the problems related to excessive state regulation, along with the interests of large
transnational corporations, led to a worldwide drive towards liberalization in the energy sector. Liberalization
itself has led to a new set of problems and challenges, which were discussed in the UNCTAD Expert Meeting
on Energy Services held in Geneva in July 2001 (see UNCTAD 2001).
16
Consumer subsidies also take various forms. In the United States, the Low Income Home Energy Assistance
Program subsidizes poor households’ fuel purchases. In many countries, especially developing countries,
consumer prices of energy are kept artificially low.
17
There is more information on energy subsidies in the US than in other OECD countries. In the US all energy
forms are subsidized, and the total amount of subsidies might reach US$30 billion (including the construction
and interest costs of the large oil reserve and the non-recoverable investment in uranium enrichment facilities).
Two-thirds are channeled towards fossil fuels, but nuclear energy is also heavily subsidized. Conversely,
subsidies to renewable energy sources are small.
21
fact even be an underestimate. The preferred forms of energy subsidies in industrialized
countries are budgetary subsidies and public provision of goods and services (including
R&D). The latter are particularly relevant in the United States and Japan (de Moor 1997).
OECD subsidy policies are biased in favour of fossil fuels, mainly coal. Total subsidies to
fossil fuels in OECD countries amount to US$ 30 billion for coal, US$ 19 billion for oil and
US$ 8 billion for gas. Adding subsidies to nuclear energy (16 billion) and to renewable
energies (9 billion), the total cost of energy subsidies per year is US$ 82 billion, equivalent to
US$ 88 per capita (van Beers and de Moor 2001, Table 3.6. See also IEA 1999).
Energy subsidies in non-OECD countries are even larger in absolute size (about US$162
billion), although lower in per capita terms (US$ 35). The bulk of these subsidies is
accounted for by ex-Soviet countries, where domestic energy prices stand at 20-40 per cent of
world prices. China and India also heavily subsidize energy, mainly electricity. Waste,
distortions and negative environmental effects are often unintended results of these subsidy
policies. Yet when considering the prospects for reforming energy subsidy measures, policy
makers face a serious policy conundrum, as energy is a crucial component of any
development strategy and access to vital energy sources is still denied to the poor majority in
many developing countries.18
Energy subsidies in many countries have been declining in the 1990s, by an amount roughly
estimated at US$ 100 billion worldwide. The trend has been particularly marked in some
non-OECD countries where energy subsidies were traditionally extremely high, notably
Russia. China, too, has been increasing coal prices up to levels close to world market ones. In
the UK, de-monopolization and privatization of e various energy sector indirectly contributed
to raise energy prices. In spite of these changes, energy is still amply subsidized in most
countries, both developed and developing.
According to OECD (1998, 2001), eliminating subsidies to energy producers in selected
OECD countries would contribute to reducing carbon dioxide (CO2) emissions (by 80 to 180
million tons by the year 2010), corresponding to a few percentage points of total emissions in
1990 but not sufficient by itself to stabilize carbon emissions. Most of the reduction would
result from the elimination of coal producer subsidies, which is expected to lead to a 20 per
cent increase in world coal prices and to a fall in total energy demand. However, if existing
non-price distortions, such as quantity restrictions (for example minimum local-purchase
obligations), were not removed as well, they would undermine the environmentally
favourable impact of the elimination of subsidies.
Phasing out subsidies would also reduce other pollutant emissions, such as greenhouse gases,
nitrogen oxide (NOx) and sulphur dioxide (SO2).
From a strictly economic viewpoint, analyses appear to show that subsidies actually tend to
damage the economy through their distorting impact, rather than strengthening it. Thus,
according to different studies, assumptions and scenarios, the net economic impact of a
18
During the Expert Meeting on energy services, the representatives of India and of various African countries
stressed the dramatic humanitarian and ecological consequences of lack of access to modern and sustainable
energy sources on the part of millions of rural and urban poor households (see UNCTAD 2001, Part.II.C).
22
mutually agreed worldwide scaling down of energy subsidies would range from negligible to
distinctively positive.19
In conclusion, evidence shows that the reduction and eventual elimination of energy subsidies
worldwide would bring large environmental, economic and even social benefits. To be
effective, fair and politically feasible, subsidy reforms should be carried out on the largest
possible scale, according to a gradual, internationally agreed and coordinated agenda.
Reforms should aim at increasing competition, and should contemplate measures to
compensate – at least in part – social groups that are bound to emerge as losers. In this
endeavour, it would be hypocritical to understate the fact that major differences separate
developed from developing countries. The latter face the harshest obstacles in reforming their
energy regimes. Conversely, developed countries – which currently consume the bulk of
world energy – are much better equipped to reform energy regimes and have a wider set of
policy options, thanks to their financial, institutional and technological superiority.
Accordingly, an asymmetric approach grounded on the principle of special and differential
treatment on the part of the international community would be a welcome and promising
prerequisite for a mutually agreed process of gradual reduction of energy subsidies
worldwide.
3.5.
Subsidies to transport services20
Transport, like energy, is essential for the functioning of any modern economy. Most
countries in the world subsidize various forms of transport subsectors, such as maritime, air
and rail transport. The bulk of the subsidies, however, is channelled towards road transport.
Subsidies to road transport are found mainly in developed countries, but they are on the rise
in developing countries (OECD 1997, 1998, van Beers and de Moor 2001). Their
conventional measurement involves an estimate of the difference between the large
government outlays on building and maintaining roads and on providing a number of services
(among them, parking spaces), as well as revenues forgone via a number of specific tax
reduction schemes (e.g. for business cars), on the one hand, and the taxes and fees effectively
collected from road users on the other hand. Many observers argue that such a methodology
leads to a severe understating of road subsidies, as it ignores the huge negative externalities
implied by this form of transportation (such as pollution, noise, accidents and congestion).
As these externalities are hard to measure, van Beers and de Moor 2001 propose a
“conventional” estimate of road subsidies that does not include them. Still, they arrive at a
total figure of no less than US$ 225 billion per year worldwide. The United States is the
country with the heaviest intervention. The country subsidizes road transport to an amount of
US$ 125 billion per year, equivalent to 2.2 per cent of GDP and to US$ 666 per vehicle.
Japan Germany, and the United Kingdom are also net subsidizers, as they earmark US$ 35
billion, US$ 13 billion and US$ 7 billion per year respectively for road transport (equivalent
19
The distribution of expected benefits, however, would be uneven, with Russia and most transition and
developing countries gaining the most, while China and energy-exporting countries either gaining less or
becoming worse off. Subsidy reforms in developing countries would also bring about economic benefits.
Studies on the social impact of energy subsidies removal tend to show that the most negatively affected group
might be the urban lower middle class, while the poor majority both in towns and in the countryside might
actually end up better off.
20
In industrialized and semi-industrialized countries, subsidies to road transport amount to indirect support to
the automobile industry. Nevertheless, road transportation as such is an activity belonging to the domain of
services.
23
to roughly one per cent of total GDP apiece). France and the Netherlands, on the contrary, tax
the sector (van Beers and de Moor 2001, table 3.9).21
Subsidies to road transport in developing countries are much lower but are on the rise,
especially in semi-industrialized countries. Governments’ stated goals are mainly to promote
overall economic development and to help the poor, but they are rarely achieved efficiently.
As in the case of water facilities, underpricing of road use can lead to lack of funds for
maintenance and ultimately to large (and avoidable) costs for premature road rehabilitation
(estimated at US$ 1.5 billion per year for developing countries as a whole in World Bank
1994). From a social point of view, moreover, subsidies to road transport tend to be
distributionally regressive. Still, as in other key infrastructural sectors, developing countries
often lack the funds and/or the technologies to set up alternative and less polluting forms of
mass transportation (such as railroads) and are in practice forced to support the growth of
road transport as the only, albeit not efficient, way to relieve major bottlenecks in
communications and economic development.
3.6.
Subsidies to R&D
The preceding subsections showed that the bulk of non-agricultural subsidies worldwide is
channelled towards three services sectors – energy, water and (mainly road) transport. There
is, however, another crucial and heavily subsidized service activity for which Van Beers and
de Moor (2001) did not present comparable aggregate data: R&D.22 This subsection presents
some indicative data on R&D expenditure and on the extent of public support to R&D
activities in OECD countries.23
Total OECD expenditure on R&D was over US$ 551 billion in 2000,24 up from less than US$
400 billion in 1990. The share of the United States was 44 per cent, and those of the EU and
Japan 28 per cent and 17 per cent respectively. In 2000, the (gross) R&D/GDP ratio25
averaged 2.24 per cent, up from 1.95 in 1981 but down from 2.29 in 1990. Sweden is the
country in the world dedicating the highest share of GDP to R&D. Among industrialized
countries, R&D/GDP ratios are above average in the United States and Japan and particularly
low in Italy and Spain. Among developing countries, the R&D/GDP ratio is very high and
increasing in the Republic of Korea and, to a lesser extent, in China. Mexico and Turkey, two
semi-industrialized OECD members, have been able to increase their R&D/GDP ratios only
modestly (OECD 2002, table 7). In most OECD countries, the share of R&D funded by the
Government26 fell in the 1990s but is still quite sizeable. It was 29 per cent in 2001, down
from 37 per cent in 1990. The respective figures for 2001 were 35 per cent for the EU and 27
per cent for the United States, showing that the weight of the public sector in R&D is
somewhat higher in Europe than in America (OECD 2002, table 9).
21
Data in van Beers and de Moor 2001, table 3.9, are taken from case studies carried out in the early 1990s.
The role of subsidies to R&D as strategic trade policy tools is discussed in subsection 2.3.
23
Systematic and comparable data on R&D expenditure are usually available only for OECD countries,
although the OECD also publishes some statistical information on R&D expenditure in other countries (such as
Russia and China).
24
OECD 2002, table 6. The figures are in constant US dollars and are expressed in purchasing power parity
terms.
25
The ratio of gross R&D expenditure to GDP (EBRD/GDP) is a useful indicator of each country’s capability to
earmark its own resources for R&D activities, relative to the size of its economy.
26
As was argued in subsection 2.3, government-funded R&D constitutes an indirect form of subsidy for a
country’s productive enterprises’ system.
22
24
The OECD also classifies R&D expenditure by performing sector: the private business
sector, higher education, government, and the private non-profit sector. The share of R&D
performed by the higher education sector increased slightly during the 1990s and stood at 17
per cent in 2001. Funding for (mostly public) higher education in the EU is relatively higher
than in the United States. The share of total R&D performed directly by OECD Governments
averages about 10 per cent. It is particularly low in the United States (7.5 per cent) and higher
in the EU (almost 14 per cent). It is also high in some developing countries: over 30 per cent
in China and Mexico. The share of the private non-profit sector is negligible (OECD 2002,
table 9). Thus, the bulk of R&D is performed by the business sector (70 per cent on average
in OECD countries, and up to 75 per cent in the United States). It has to be taken into account
that part of private R&D is also government-subsidized and may benefit from fiscal
incentives (tables 3.2 and 3.3).
Most OECD countries are presently changing their S&T policies in a more proactive sense,
after a decade of fiscal restraint and stagnation, and they are increasing public funding for
R&D and innovation. Some countries have particularly ambitious targets, among them
Austria, Canada, Norway and Spain. The EU has set a goal of earmarking 3 per cent GDP for
R&D by 2010. The Republic of Korea is determined to increase government R&D
expenditure to 5 per cent of the total government budget.
Government efforts to promote industrial R&D and innovation are being strengthened, a
number of reforms are being implemented in universities and public research organizations,
and the focus of research on selected targets is being sharpened in order to maximize R&D’s
positive impact on economic development (OECD 2002, pp.53-54). Government support
programmes for private firms’ research efforts are being stepped up in most OECD countries,
many of them targeting SMEs. Among the countries that have recently established new
publicly funded R&D support programmes, or have been strengthening existing ones, are
Australia, Canada, Germany, Hungary, Ireland, the Netherlands, New Zealand, Portugal,
Spain, the United Kingdom and the United States.
Table 3.2
Industry R&D expenditures and government financing
(Mid- to late 1990s)
Total EBRD
Government financing*
(Bn of 1995 PPP US$)
(Percentage of total
EBRD)
Australia
Canada
France
Netherlands
United States
3.2
5.1
14.2
3.2
152.6
6.9
22
15.1
12.5
17
Source: OECD 2002
Note: BERD European bank for reconstruction and development
*Sum of cost to Government of tax credits plus direct government funding.
In 2000 and 2002, many countries also modified their tax treatment of private R&D in order
to make it more business-friendly, enhancing its incentive/creative potential, and increasing
effective tax reduction rates (OECD 2002, pp.65-67). Other programmes support
25
entrepreneurship and growth of SMEs by reducing red tape, promoting start-ups and
innovative firms, stimulating R&D and innovation in SMEs, enhancing cooperation and
technology diffusion among public and private firms and research organizations, and
facilitating the commercialization of public research. Some of these programmes are being
carried out with a cluster-based or regional innovation focus.
Other closely related policy efforts are aimed at strengthening OECD countries’ scientific
human capital by increasing public support for key sectors of higher education in order to
increase the quantity and quality of scientists and engineers (OECD 2002, pp.64-67).
Table 3.3
R&D Tax subsidies in manufacturing companies, 2001
(Per US$ of R&D)
Canada
France
Germany
Japan
Korea, Rep. of
Mexico
UK
US
Small firms
0.322
0.061
-0.025
0.121
0.111
0.031
0.106
0.066
Large firms
0.173
0.061
-0.025
0.009
0.126
0.031
0.096
0.066
Source: OECD 2002
4.
SUBSIDIES IN MULTILATERAL TRADE NEGOTIATIONS
4.1.
Disputes on subsidies in the early GATT years
Negotiations on subsidies have posed and still pose “particularly difficult problems for the
multilateral system because (they challenge) state sovereignty and core values represented by
individual states…the major lesson gleaned from a study of the multilateral subsidy issue is
the difficulty of reconciling domestic imperatives with international pressure for change”
(O’Brien 1997, p.102).
In the immediate post-war years, multilateral trade diplomacy followed two paths: a “fast”
one and a “regular” one. The fast path was that of the GATT, which was originally a limited
tariff-cutting agreement, signed by the United States and another 14 countries in 1947. The
“regular” path was more ambitious and universal in nature, and was based on the Havana
Charter (1948), which envisaged a wider programme of trade liberalization, including the
elimination of export subsidies. The Charter should have paved the way for the birth of the
International Trade Organization (ITO), into which GATT was expected to be subsumed.
However, serious contradictions – mainly centred on the issue of export subsidies – erupted
the between the US State Department, a strong herald of free trade, and the US Agriculture
Department, which was adamant about protecting US agriculture. As a result, the United
States did not to ratify the Havana Charter, thus effectively killing ITO before it could even
come into existence. GATT, which was originally meant to be a temporary and partial
26
instrument, “became permanent and was left as the main legal framework governing
international trade” (O’Brien 1997, p.103).
With respect to subsidies in particular, it is interesting to note that the degree of liberalization
in the domain of subsidies envisaged by the GATT in its original formulation was much
lower than that foreseen by the Havana Charter. In the GATT, Article XVI handles subsidy
practices, while Article VI focuses on countervailing duties. Article XVI did not explicitly
mention export subsidies, stating only that countries granting any kind of subsidies, which
might impact on international trade had to notify them to GATT and consult bilaterally with
other countries that might be negatively affected. Initially, disputes focused mainly on the
consequences of subsidies on export competitiveness in third countries, but the practical
impact of GATT on subsidizing policies worldwide was minimal.
Increasing realization of the limited effectiveness of Article XVI led to its revision in 1955.
Subsidies were divided into various subcategories—domestic, export, primary, nonprimary—to be treated in different ways. Export subsidies were disciplined more severely
than domestic subsidies, and subsidies to industry more severely than subsidies to the
primary sector. A 1960 declaration further strengthened this differential treatment,
prohibiting subsidies aimed solely or primarily at supporting industrial exports. Implicitly,
GATT rules accepted that most subsidies were granted primarily for domestic purposes, and
it would not have been politically feasible to eliminate them altogether.
Clearly, its reliance on the subsidizing country’s initiative weakened the effectiveness of
Article XVI. Countries that deemed themselves to be damaged by other countries’ subsidies
could, however, levy a countervailing duty (CVD) in accordance with Article VI. Article VI
conditioned the imposition of CVDs on an injury test (i.e. it had to be demonstrated that the
subsidy was inflicting significant damage on the requesting country’s industry). However,
thanks (formally) to a technical legal loophole, the United States, the only country to actively
use CVDs, managed to be exempted from applying the test. This exemption left the United
States free to use this measure as a protectionist tool, but CVDs were in fact applied much
less extensively than antidumping duties. In this respect, it has to be pointed out that the two
instruments are only apparently equivalent: while antidumping duties target unfair trade
practices of foreign firms, CVDs directly challenge the policies of another State and thus its
very sovereignty. The political sensitivity of CVDs is therefore higher than that of
antidumping duties, a fact that largely explains the wider popularity of the latter.
4.2.
Subsidies in the Tokyo and the Uruguay Rounds
The issue of subsidies did not provoke strong disputes until the 1970s, when, as a result of the
Tokyo Round negotiations, a subsidy code was added to the GATT and CVD rules were
tightened. During the negotiations, the United States adopted an aggressive position, arguing:
“Subsidies are serious trade distortions that must precede countervailing action. Therefore,
why should the onus be on the countervailer rather than the subsidizer?” (Winham, 1986,
p.119). Subsidies, moreover, distorted trade simultaneously in three markets: if country A
subsidizes its industry x, it is likely to import less x products from country B, it will export
more quantities of x to country B, and will unfairly boost its share in country C’s market,
crowding out B’s exports of x. The EC and most other countries, conversely, saw the US use
of CVDs as the most relevant danger to trade relations, as they were like a Damocles’ sword
threatening to shut off access to the world’s largest market.
27
During the Tokyo Round negotiations, which were finalized in 1979, the United States
proposed (for the first time) a traffic light approach to the subsidy issue. Eventually a Subsidy
Code was adopted, stating that the United States would go through an injury test to discipline
its use of CVDs, and other countries would curb their subsidy practices. The results of the
Code were unsatisfactory. On the one hand, lobbying from industrial interests led Congress to
drain the injury test of any significance, and the use of CVD measures on the part of the
United States actually rose dramatically in the early 1980s. On the other hand, various
factors—such as problems with the definition of subsidies, clauses acknowledging the
legitimacy of subsidies to achieve a wide gamut of domestic social and economic goals, and
the absence of an effective multilateral mechanism (as distinct from the unilateral imposition
of CVDs) to sanction unfair subsidy practices—contributed to weakening the impact of the
Code on the widespread implementation of subsidy policies all over the world.
Accordingly, pressure from different sides to deal more effectively with the subsidy issue in
the Uruguay Round negotiations, which had started in 1986, intensified. Disputes about
subsidies were at the core of the negotiations and went through a complex series of ups and
downs, due not only to the traditional, structural divergences between the United States and
other OECD members and between developed and developing countries, but also to the
change in the US position itself, which during the Clinton administration became more
favourable to industrial policy interventions, especially in the domain of high technology
sectors. The Uruguay Round Agreement was concluded by the end of 1993, and the subsidy
issue was regulated through the traffic light approach. Basically, non-actionable subsidies
(green) are those that are generally available rather than being specific to particular
industries, those granted to favour the development of disadvantaged regions and—under
more restrictive conditions—subsidies to enterprises investing in new machines and facilities
in order to meet higher environmental standards. Prohibited subsidies (red) are export
subsidies and subsidies conditional on local requirements. All remaining subsidies that are in
law or in practice specific (i.e. available only to a single firm or industry, or a determined
group of industries) are actionable (yellow or amber). This means that a country that
considers itself to be negatively affected by another country’s subsidy policy can impose
CVDs, provided a number of conditions are met (among them the existence of serious
prejudice for domestic industry as a whole and a level of subsidy exceeding a de minimis
level).
Bargaining among the various countries was particularly lively in the domain of R&D
subsidies—with the United States now pushing for an even higher upper permissible limit
than the Europeans—and with respect to the issue of the specificity of subsidies earmarked
for particular subnational geographic areas. The latter dispute was resolved with a decision
that generally available subsidies in a determined subnational area would not automatically
be considered specific.27 This point was far from being irrelevant in practice, especially for
the United States, where individual states implement arrays of subsidy programmes worth
billions of dollars, many of which are earmarked for services sectors and activities such as
energy, audiovisuals and R&D—as detailed and thoroughly documented in two 2002
27
The fact that all subnational subsidy policies are not automatically deemed specific does not imply that a
particular subnational subsidy programme can be considered specific by a trade rival, who might try to prove its
case in order to impose CVDs. There is in fact ample room for future trade disputes focusing on the issue of
specificity of regional or state-level subsidy policies.
28
notifications to the WTO Committee on Subsidies and Countervailing Measures (SCM 2002a
and SCM 2002b).
Sectoral issues, centring on agriculture, aerospace and audiovisual industries, were also hotly
debated. In the final WTO agreement, the discipline on agricultural subsidies, like all other
aspect of trade in agricultural products, was regulated in a unique, ad hoc fashion, while
aerospace was not brought into GATT at all. As the focus of the present paper is on subsidies
to services, rather than goods-producing sectors, the subsidy rules in these two sectors are not
examined in detail, and only one crucial point needs to be mentioned. Generally speaking,
room for policy intervention, including subsidy programmes, is much larger for agriculture
and aerospace than for most other sectors. This also means that opportunities for countries to
subsidize infrastructural and other services (including the most important one, R&D), the
contribution of which is increasingly critical in determining international competitiveness, are
much wider in agriculture and aerospace than in other sectors. It is plain that such an
asymmetry, for agriculture in particular, exacerbates the overall distortions implied by the
exceptionally high level of interventionism allowed by existent, sector-specific WTO
regulations. Such distortions work perversely against the poorest countries, which are not in a
position to engage in massive investment in R&D and other agricultural-productivityenhancing services and therefore see the potential competitive advantage stemming from low
rural labour costs dramatically eroded.
The audiovisual sector, on the other hand, is a service sector in its own right (although some
audiovisual products are also traded as commodities). It is therefore essentially regulated by
GATS, not by the GATT. GATS, unlike the GATT, leaves ample room for the
implementation of various forms of subsidy policies. Moreover, it contains specific
provisions for the audiovisual sector, allowing Governments virtually a free hand in their
subsidy policy options. In the Uruguay Round, however, this outcome was far from obvious,
as the United States pushed hard to achieve a significant degree of liberalization in this
sector. However, opposition from other countries, and from European countries in particular,
which adamantly defended the cultural exception principle, proved too strong to be
overcome, and the prospect of a liberalization of trade in audiovisuals, as in other services,
was left to future rounds of GATS negotiations (Box 4.3).
In short, the Uruguay Round agreement on subsidies and CVDs represented a substantial step
in the direction of multilateral trade liberalization. Yet the disputes that immediately erupted
in the US Congress when it started debating the ratification and implementation of the
agreement were a signal that the controversy was not over. As previously pointed out,
“subsidy negotiations are difficult because they increase the possibility of exposing domestic
policy to severe international constraints” (O’Brien 1997, p.125). Room for jockeying with
the interpretation of the norms governing the applicability of CVDs, the specificity of
subsidies and the scope of subnational support programmes to score new points against
trading rivals is ample, especially for stronger countries. The likelihood of the emergence of
disputes on subsidy policies is increased, moreover, by the worldwide deceleration of
economic growth, with its unequal impact on various countries, industries and sectors, and by
the break-up of the previously quasi-unanimous consensus on the desirability of universal and
ever-deepening trade and financial liberalization.
29
4.3.
The Agreement on Subsidies and Countervailing Measures and special and
differential treatment for developing countries
References to services are numerous in both the GATT and GATS, but in the Uruguay Round
the issue of services was treated specifically in an additional instrument, the Agreement on
Subsidies and Countervailing Measures (ASCM). For the time being, the ASCM only applies
to industrial subsidies, as both agriculture and services were basically left out (subsection
4.4). However, its contents are not irrelevant to the focus of the present paper, as its
principles are likely to constitute a key reference point for present and future negotiations on
the issue of subsidies in the domain of trade in services as well. Part I of the Agreement states
that the ASCM only applies to subsidies provided specifically to an enterprise or industry or
to a group of enterprises or industries. Parts II, III and IV set out the “traffic light approach”.
Part V describes the substantive and procedural requirements a WTO member must fulfil
before applying any countervailing measure. Parts VI and VII provide the institutional
structure and the notification and surveillance modalities. Part VIII contains special and
differential treatment (SDT) rules for developing countries. Part IX establishes transition
rules for developed and transition countries. Parts X and XI detail the dispute settlement and
final provisions (UNCTAD 2000). This subsection focuses on the issue of SDT.
The agreement recognizes that “subsidies may play an important role in economic
development programmes of developing country members”. Thus, it accords exemptions
from the restrictions on subsidies to three categories of developing countries: LDCs,
countries with a per capita GDP of less than US$ 1,000 per year, and other developing
countries. The poorer the country, the more favourable the treatment. The first two categories
of countries are exempted from prohibition, while the other developing countries can
eliminate subsidies over an eight-year period. During that period, they cannot increase export
subsidies, and they must eliminate them earlier if the subsidies are not consistent with their
development needs.28 Developing countries that consider it necessary to apply export
subsidies beyond the eight-year period might be allowed to extend them over a limited period
of time after consultation with the SCM Committee. With respect to import-substitution
subsidies, LDCs are given eight years to eliminate them, and countries with a per capita GDP
of less than US$ 1,000 per year are given five years. Developing countries are also treated
more favourably than developed countries in the domain of actionable subsidies,
countervailing measures and dispute settlement.
The practical impact of the SDT dispositions set out in Part VIII is meagre, taking into
account the safe haven created in the advanced, supply-chain management for several
subsidies used intensively by developed countries, which were made non-actionable (mainly
R&D, regional development subsidies and environmental subsidies). On the contrary, as
previously mentioned, “subsidies used by developing countries for development,
diversification and upgrading are actionable since action can be taken against them under
certain conditions. This represents a grave imbalance in the agreement” (UNCTAD 2000,
28
An additional principle applies, related to the concept of “export competitiveness”. Generally speaking, the
phasing out of export subsidies is expected to take place in a shorter period of time if a developing country
achieves a state of “export competitiveness”, defined as a certain share of world trade for a given product for
two consecutive years. Such a share is normally 3.25 per cent, but additional specific norms further condition
the applicability of the export competitiveness principle, discriminating again in favor of the poorer developing
countries (UNCTAD 2000, pp. 9-10).
30
p.22). Moreover, the policy room that developing countries are allowed to apply importsubstituting subsidies is constrained, and might be challenged as inconsistent with TRIMs.
The export competitiveness principle eventually nullifies any advantages of developing
countries in the domain of subsidies at the very stage when a developing country acquires a
non-negligible position in world trade.29 By the same token, developing countries
automatically lose most of the SDT benefits as soon as they reach a per capita GDP of US$
1,000, a low threshold indeed. A number of dispositions on agricultural subsidies and dispute
settlement also fail to support the spirit of the SDT principle (UNCTAD 2000, section V).
4.4.
The treatment of subsidies in the GATS
One major result of the Uruguay Round was the creation of a powerful, high-level
multilateral international trade body, the WTO, which supersedes the implementation not
only of the GATT (which currently regulates international trade in goods) but also of GATS,
a new agreement specifically tailored to regulate international trade in services. Taking into
account that service sectors worldwide tend to be much more regulated than goods sectors, as
well as a number of other differences between trade in services and trade in goods, it can be
said that the coming to existence of GATS effectively removed the issue of subsidies to
service activities from the ASCM discipline.
Due to a complex web of factors (among which it might be worth mentioning the fact that
trade in services is effected through four modes of delivery, as well as other structural
differences between trade in services and trade in goods; the traditionally higher degree of
state intervention, and the lower degree of tradability of services, with respect to industry and
agriculture; the statistical and monitoring difficulties inherent in the estimation of
international trade flows in services; and the negotiating stances of major trading partners
during the Uruguay Round negotiations), the extent of binding trade liberalization brought
about by GATS in the domain of international trade in services falls short of that brought
about by GATT 1994 in the domain of trade in goods. With respect to the issue of subsidies
in international trade in services, it can be broadly stated that GATS does not restrict in any
practical way member States’ faculty to implement subsidy policies.30
Subsidies are dealt with specifically in Article XV of GATS. In fact, however, Article XV
does not go any further than recognizing that “in certain circumstance, subsidies may have
distorting effects on trade in services” and advocating further negotiations on the issue
(UNCTAD 1993). To date, therefore, subsidies for services are fully GATS-compatible.
Subsidies, however, should not in principle prevent the granting of MFN treatment, and
hence should not discriminate in favour of national service suppliers. It should also be
noticed that nothing in GATS prevents countries from increasing their subsidies.
Other articles can be invoked to justify the enactment of subsidy policies in a more indirect
fashion. Article I, in subsection 3, states that GATS “includes any service in any sector
except services supplied in the exercise of government authority“ and that the latter are all
those services “supplied neither on a commercial basis, nor in competition with one or more
service suppliers”. Along with Article XIII, this subsection leaves member countries room for
29
An additional problem with export competitiveness might be constituted by its asymmetry: if a developing
country loses export competitiveness, it is not clear how it might be allowed to apply export subsidies again
(UNCTAD 2000, para V (iii)).
30
The picture might change, of course, if the present round of GATS negotiations produces a consensus to
change the GATS structure substantially in the direction of progressive but forceful liberalization.
31
several kinds of “proactive” policy interventions to favour national producers, for instance in
the audiovisual sector, using subsidies or other policy instruments (Box 4.3). Article XIII
rules that Article II and XVII (on most favoured nation treatment) and XVI (on market
access) do not apply to the “procurement by governmental agencies of services purchased for
governmental purposes” in the pursuit of non-commercial goals. Of course, government
procurement can be a form of subsidy policy.
4.5.
Multilateral regulation of subsidies to service activities: a grey area?
One of the key requisites for subsidies to be actionable (or totally prohibited) in the ASCM is
their specificity (subsection 3.3). Because of the peculiar characteristics of services, however,
the attribution of specificity is more difficult than in the case of industry or agriculture, as
many services have an infrastructural nature and contribute indirectly to the functioning of
many other sectors or even of the economy as a whole. Total or partial provision of services
by the Government—often justified primarily for social reasons—is widespread. Various
forms of indirect subsidization (among them) also affect the international competitiveness of
firms, blurring the separation between trade in services and trade in goods.31 Particularly
widespread forms of indirect subsidization include the preferential tax regimes for foreign
operations and foreign consumers, and the subsidies to ancillary service activities needed to
support productive sectors, such as R&D, travel and feasibility studies.
Only some of these non-trade-neutral subsidies to services sectors are clearly and specifically
regulated by multilateral trade rules. An example is subsidies to export shipments provided or
mandated by Governments on terms more favourable than for domestic transportation, which
are prohibited under GATT Article XVI. Conversely, subsidies to consultancy services are
not only allowed but are non-actionable. Other input and infrastructural services (such as the
preferential provision of electricity at lower rates and government-financed construction of
storage, processing and transport facilities, to the extent that they are targeted specifically at
exporters) are in principle actionable, and in fact have been the cause of WTO disputes and
the imposition of countervailing measures—especially on the part of the United States
(UNCTAD 1993, pp. 4-5). The complexity of the matter and the lack of clear-cut principles
on the discipline of subsidies in trade in services, however, have made and continue to make
these cases highly controversial.
The uncertainties surrounding multilateral regulation of subsidies for infrastructural services,
in particular, allow the existence of a grey area, as many of these services are in fact
subsidized, thus altering the international competitive playing field in a number of important
traded activities. Transport and utility services are cases in point. Among the former, air and
maritime transportation have been the object of various trade disputes, a fact that reserved
them special treatment in the Uruguay Round agreement.32
31
Subsidies to a service activity can be enacted through a number of policy tools, among them the subsidization
of goods used as inputs in producing the services (examples include ships, aircrafts and hotels) and government
purchases.
32
Subsection 4.5 – the magnitude of subsidies to transport services.
32
Financial and audiovisual services, of course, are also the object of heavy public intervention
in most countries. The impact of financial policies, including preferential lending
programmes, on the economy as a whole, and on export competitiveness in particular, is selfevident. Audiovisual services, too, tend to be subsidized and protected (Box 4.3).
Other infrastructural or social services, such as education and health, are often provided by
Governments at little or no cost for obvious social reasons, yet they do have a strong systemic
impact on international competitiveness (see subsection 4.6 on subsidies to R&D and higher
education). Besides their indirect impact on the economy as a whole, some (often subsidized)
education and health services can also be directly exported. This is the case in many
developed and also in a few developing countries (the United Kingdom and the United States
for education services, and Cuba, India, Switzerland and the United States for health
services). In spite of the existence of specific multilateral trade rules, potentially tradedistorting subsidies for the production and consumption of telecommunication services—a
sector that has shifted from public to private hands in a number of countries in the last two
decades—are also common (UNCTAD 1993, pp.6-7).
Another category of subsidies to service activities that can be considered to amount to
indirect export subsidies comprises subsidies to “export market development” programmes,
which cover several activities aimed at the penetration of new markets and the preparation of
feasibility studies, projects, bids and the like. On one hand, government support for such
activities can be a badly needed and very effective policy tool for developing countries in
overcoming some of their firms’ structural weaknesses and helping them to get a foot in
international markets. On the other hand, given the asymmetry of Governments’ financial
capabilities, subsidies for export market development more often than not end up being an
additional factor magnifying the competitiveness gap between firms from developed and
developing countries. This is the case, for instance, for engineering and construction services,
as was documented by a number of examples in the UNCTAD Expert Meeting held in
Geneva in October 2000 (UNCTAD 2000a).
In short, subsidies for social and infrastructural services, even more than other types of
subsidies, do have an impact on international competitiveness. So far, due among other things
to the weakness and unsystematic character of multilateral disciplines on this matter,33 these
subsidies have tended mostly to favour developed countries’ trade competitiveness at the
expense of developing countries. If the special and differential treatment principle were more
generally accepted and more effectively applied in this domain, however, developing
countries’ trade and development prospects could be significantly boosted.
4.6.
Analytical work on subsidies to sector services in the WTO: the Working Party
on GATS Rules
After the conclusion of the Uruguay Round and the launching of GATS, a number of pending
issues in the domain of multilateral rules on trade in services were left to further negotiations.
Subsidies, in particular, were to be discussed in the Working Party on GATS Rules.
Responding to a request by the Working Party, the WTO Secretariat, in January 1998,
33
As mentioned in other parts of this paper, there are a number of objective factors that make the regulation of
subsidies more complex and difficult in the case of services than in the case of goods. It has been argued, for
instance, that countervailing duties are more difficult to apply in a proper fashion and could more easily turn out
to be disproportionate and arbitrary in the case of trade in services than in the case of trade in goods (UNCTAD
1993, p.11).
33
presented a note (WTO 1998) containing empirical evidence on subsidy programmes for
service sectors in various countries and larger trading entities (such as the EU), based on all
the WTO Trade Policy Reviews (TPR) conducted since 1995 (31 in total).
The definition of subsidy used in the note is basically the same as that established by the
WTO Agreement on Subsidies and Countervailing Measures, which states that subsidies are
financial contributions by Governments (or other public bodies) that confer a benefit.
Consistently, TPR reports cover support to service activities bestowed “in the form of direct
transfers of funds, including grants, loans, and equity infusions; potential direct funding of
funds or liabilities, e.g. loan guarantees; government revenue forgone; supply of goods and
services other than general infrastructure; purchase of goods; payments to funding
mechanisms, or income and price support” (p.2). Conversely, it does not include other, more
indirect forms of subsidization, such as: “exemptions from indirect taxes (i.e. VAT, usually
aimed chiefly at encouraging consumption rather than production of a service); generally
available support (i.e. in the context of regional development or R&D programs); companyinternal cross-subsidization between monopoly and market-oriented activities in sectors such
as telecommunications” (WTO 2000b, p.2).
With respect to this definition, the note observes that it is correct from a conceptual point of
view, but the main problem it raises consists in identifying the ultimate beneficiary of a
subsidy in a given services area, as the support measure may actually have been earmarked
for downstream users, rather than the immediate recipient. In addition, subsidies to service
activities are often aimed primarily at public policy or infrastructural objectives. An example
is provided by the provision of basic health services, which can be ensured through different
measures such as cost-free treatment in state-owned hospitals (a public infrastructural
service), the granting of public funds to private hospitals (a form of subsidy to the health
services sector), or government-sponsored premiums for basic health insurance (a form of
subsidy to the insurance services sector).
In the introduction, the WTO Secretariat gives a warning on a number of issues. The note
focuses only on one specific policy instrument—financial support—“whose use may be
accompanied by a variety of other measures, such as regulated prices and access conditions”
(p.1). The TPR reports could not always identify thoroughly the reciprocal interactions
among these measures, or evaluate their relative importance in a comprehensive fashion. As
mentioned above, the very goals of the numerous subsidy programmes were often difficult to
disentangle, as many of them were likely to have been implemented mainly for non-trade
purposes, in accordance with social or industrial policy objectives. Moreover, “given the
infrastructural importance of many services sectors and their role as generally available
inputs, it could be misleading in individual cases to equate the recipients, e.g. railways, with
actual beneficiaries which may include any social or economic group relying on rail
transport” (ibid.). The selection of countries for review did not follow any global scientific
and statistical method, but merely followed the schedule previously established by the Trade
Policy Review Body. Finally, the coverage and contents of the reports were determined by
the availability of data, which in turn was largely (albeit not exclusively, as the TPR reports
also strived to include relevant information from other sources, such as international
organizations or independent researchers) a function of the Government’s ability and
34
willingness to provide the Secretariat with adequate cooperation.34 For all these reasons, the
Secretariat acknowledged that the note could “not claim to give a representative picture”.
Still, it constitutes a “reasonably comprehensive source of information in this area”(ibid.).35
To sum up, from a methodological viewpoint, the note is in a way more significant for what it
does not say than for what it says. Still, it does propose a few brief remarks stemming from
the comparative examination of TPR reports’data.
Box 4.1
An interpretation of the law of subsidies
Due to the complexity and the sensitivity of the issue and to the progressive superposition of various
and not always obviously consistent agreements and amendments, it might appear that multilateral rules on
subsidies in the GATT/WTO system in fact constitute a legal patchwork. On the contrary, according to Benitah,
there is a “fundamental theme” underlying all provisions on services in the system and which “constitutes the
key to understanding their legal basis”. This theme confronts the need to attenuate entitlements granted to the
party seeking to defend itself against the “adverse effects of subsidies” (Benitah 2001, p.1). The legal treatment
of subsidies bears a resemblance to that of pollution: the negatively affected party sees both pollution and
subsidies as negative externalities, and seeks complete protection from the legal system. However, the modern
economy could not function if every kind of polluting economic activity were fully prohibited. Thus, legislators
often deem that the economic and social cost of eliminating the polluting activity completely would be higher
than the environmental benefits. Moreover, they realistically realize that the functioning of the political
mechanisms would not in practice let such a draconian piece of legislation go through. Thus, they usually end
up producing a balanced legal instrument, which confers the negatively affected party only partial protection,
based not on administrative prohibition but on a stick-and-carrot approach (i.e. stipulating that the polluting firm
will pay a compensation fee to specific, negatively affected private agents, and/or a tax to the community as a
whole, and that these financial onuses will increase proportionally or progressively with the level of pollution).
This approach is consistent with the classical Coasian approach to social choice: “We are dealing with a
problem of a reciprocal nature…should A be allowed to harm B or should B be allowed to harm A? The
problem is to avoid the more serious harm” (Coase 1960, p.2).
By the same token, GATT/WTO negotiators soon realized that a wanton prohibition on any kind of
subsidy was neither economically nor socially and politically feasible. They therefore progressively created a
multilateral legal system that provides only partial protection to the party subjected to the “adverse effects of
subsidies. In this endeavour, according to Benitah’s interpretation, negotiators utilized legal techniques of
“attenuation” of the negatively affected party’s entitlements: “it will be rare to find a situation where a practice
is completely prohibited. One will generally find a technique which consists of permitting a practice but
subjecting it to the obligation of avoiding certain effects” (p.18). As a result: “Each time an entitlement is
granted to this party, we…note the presence of legal techniques intended to attenuate its scope…(although) the
systematic use of such techniques…(presents) inherent dangers…the fact of having recourse to these techniques
of attenuation goes hand in hand with the need for their clarification through a case law process”(pp.1-2).
34
This asymmetry might cause a selection bias, if two different TPR reports were superficially compared.
Paradoxically, country A, after having fully and exhaustively cooperated with the WTO Secretariat, might end
up being penalized as excessively protectionist, while country B, which could not or did not wish
to reveal the real extent of its subsidy programmes, would wrongly appear to be a virtuous free trader.
35
As the WTO is the highest and—by now—virtually universal authority in the domain of international trade,
the numerous and serious caveats recognized by the authors of the note indirectly suggest that there are vast and
serious limitations on any independent researcher’s ability to comprehensively assess and quantify the pervasive
phenomenon of the subsidization of service activities.
35
Subsidies appeared to be concentrated in four sectors: audiovisual services, air and maritime
transport, tourism, and banking.36 Subsidy programmes for audiovisual services were
particularly relevant in developed countries, while subsidies to the tourism sector were
enacted mainly in developing countries. They usually relied on tax holidays and other fiscal
advantages, rather than on actual disbursement of public funds. Subsidies to transport and
financial services (the latter often being of an emergency, ad hoc nature) were common in all
types of countries.
In May, 2000, the WTO secretariat issued an addendum (WTO 2000a), which updated the
contents of the 1998 note, utilizing information extracted from another 37 TPR reports
published in the meantime. In the new sample, the policy patterns did not appear to be very
different from those outlined two years previously. The three most subsidized services were
still among those identified in the 1998 note: maritime transport, tourism and financial
services. Tourism subsidy programmes, identified in 22 out of 37 countries, were particularly
popular in developing countries. Audiovisual services and airlines, however, appeared to
have lost some relative weigh. A new trend appeared to be identifiable in the financial
domain: besides emergency rescue interventions, strategic programmes aimed at developing
the sector through tax incentives and/or offshore schemes were also widespread.
Box 4.2
The definition of R&D subsidies and the US negotiating position
During Uruguay Round negotiations on subsidies, the United States, in line with its traditional liberal
philosophy, strove to expand the list of prohibited (red) subsidies. Yet it met strong resistance, especially on the
part of the EU, which bargained hard to maintain a regime that explicitly allowed certain categories of subsidies,
among them those for R&D. Negotiations on R&D subsidies centred around three sets of questions: the first set
was definitional, the second was sectoral, and the third concerned how existing assistance programmes
(including those being enacted in the United States) would be affected by the new multilateral regime.
Definitional questions focused on making explicit precisely what lay behind the acronym “R&D”.
What is research? What is the difference between applied, basic, and experimental research? What is
development? The US Advisory Committee for Trade Policy and Negotiations, in a 1990 report, argued that
subsidies should be allowed for basic and experimental research and (with more limitations) for research
projects in general, but that a harsher approach should be adopted for subsidies to development. Yet members of
the Committee were themselves aware that to distinguish between the various phases of research and
development is often not easy, and that depending on the sector or even on the specific project, the criteria to
assess the degree of “basicness” of a specific activity might vary to a large extent, besides being perceived
differently by the various stakeholders. A particularly delicate issue was the interpretation of footnote 29 of the
Subsidies Agreement, which referred to the term “pre-competitive development activity”, which would include
the creation of a “first prototype which would not be capable of commercial use”(quoted in Morris 1998, p.3).
Thus, further studies were recommended, but no agreed clear-cut conclusions were reached. US
negotiators were quite sympathetic to the plight of American civil aircraft constructors, who complained about
what they considered unfair competition on the part of heavily subsidized Airbus. However, they also gradually
realized that excessively harsh regulation of subsidies could eventually jeopardize the survival of numerous
large programmes already implemented in the United States. US trade representatives understood that the
policy-supported cooperation between government and advanced industry was at the very core of the
technological and economic dynamism of American society, and they feared that the mere process of complying
with the publicity and transparency requisites foreseen by the WTO machinery would end up leaking strategic
information to foreign competitors (Morris 1998, p.4).
36
Out of 44 countries reviewed, 17 subsidized their audiovisual services, 14 their hotel and tourism sector, 13
their national airlines and many their maritime transport fleet. 18 countries assisted their banking sector.
36
A second addendum was issued by the end of 2000 (WTO 2000b). It took stock of the results
of nine new TPRs. The new document confirmed the policy patterns previously observed:
WTO members concentrate subsidies on tourism, maritime transport, and banking services.
For clear political and fiscal feasibility reasons, foregone revenues in the form of tax
concessions were more popular than direct grants.
In September 2002, the WTO issued a third addendum (WTO 2002), covering 24 new TPRs.
In the introduction, the Secretariat noted that, as in previous similar exercises, the operational
definition of subsidies to the service sectors included Governments’ share purchases and
similar equity transfers.37 It also reiterated the problems often encountered in identifying
subsidies’ ultimate beneficiaries, taking into account that subsidies to many infrastructural
services indirectly benefit all downstream users. Other problems stemmed from the fact that
some subsidies are in fact targeted at services provided “in the exercise of governmental
authority” (GATS Article 1:3(b)) and from the difficulty, in some cases, of ascertaining if the
subsidized product is actually a service or a good. Even the sectoral policy patterns observed
might to some extent stem simply from the fact that TPR reports tend to focus on tourism,
maritime transport and banking. With all these caveats, WTO 2002 confirmed the main
features of previously identified policy patterns in subsidizing services.
Box 4.3
Can Governments implement subsidy policies to support audiovisual services,
according to multilateral trade discipline?
Due to their unique nature, audiovisual services deserve very special treatment in international trade
discipline. Yet the “cultural exception” is not mentioned as such in major multilateral trade agreements,
although they do contain specific provisions establishing special treatment for some forms of trade in all or
some cultural goods and services (UNCTAD 2002). Actually, even the attribution of sovereignty on the matter
of international trade in audiovisual services can be disputed. In fact, while there is no discussion on the full
attribution to the GATT of jurisdiction on trade in most ordinary goods, opinions differ as to whether
jurisdiction on audiovisual services should essentially be a prerogative of WTO or UNESCO or should
somehow be shared by the two international organizations. The dispute is not purely academic. “WTO’s
orientation is towards the economic side of any issue…(while) UNESCO…adopted an increasingly sceptical
approach to the role of the market in the production and diffusion of cultural goods and services (van Grassteck
2001).
From the point of view of comprehensive international legality, the attribution to either organization of
clear-cut competencies on audiovisual services is not uncontroversial either. For instance, some UNESCO
agreements on trade in cultural property, such as the 1970 Convention on the Means of Prohibiting and
Preventing the Illicit Import, Export and Transfer of Ownership of Cultural Property, while only secondarily
trade-related and not obviously in contrast with WTO provisions, might in fact constitute grounds to restrict
trade in cultural goods and services and hence also in audiovisual services. More recently, UNESCO has been
emphasizing the need to promote “cultural diversity”, combating the risks of worldwide cultural standardization.
In this endeavour, UNESCO argues that cultural goods and services require differential treatment in multilateral
trade discipline and a strong regulatory framework to support national cultural policies. From the vantage point
of international trade proper, therefore, it is conceivable that future negotiations, and quite conceivably actual
trade disputes on trade in audiovisual services, will see one or more of the parties involved invoking the
alternative jurisdiction of UNESCO as opposed to that of the WTO. The non-exclusivity, at least in principle, of
the WTO role in regulating trade in audiovisual services can actually be seen as a broad implication of the
relevance of the “cultural exception” principle. In this respect, it is important to remember that subsidies are by
far the most widely used instrument used by Governments—in both developed and in developing countries—to
37
Such an inclusion had been criticized on the ground that Governments, in some cases, might mainly have
targeted macroeconomic, rather than sectoral objectives. Yet, for the sake of consistency, the Secretariat deemed
it preferable to continue applying the same methodology,.
37
support the audiovisual industry. Bringing UNESCO into the picture is likely to strengthen the hand of
negotiating parties opposing the introduction of new binding limitations to national States’ flexibility in
implementing subsidy policy in the audiovisual sector. Another two important trade agreements negotiated in
the Uruguay Ground, namely the GATT and the TRIPS, also have some implications for trade in audiovisual
services. GATT, in particular, contains several exceptions inspired by the same need to preserve ample room for
Governments to pursue their own cultural policies and to protect national audiovisual industries (Article IV,
Article XIX, Article XX). However, neither GATT nor TRIPS norms specifically affect the acceptability of
subsidy policies for supporting audiovisual industries.
Essentially, however—with the caveat represented by the uncertainties surrounding the role of
UNESCO, the practical implications of which are difficult to foresee—international trade in audiovisual
services, like trade in any other services, is presently disciplined by the GATS agreement. GATS, besides
allowing the use of subsidies in services in general terms, contains various dispositions which—directly or
indirectly—further widen the policy room for would-be subsidizers of audiovisuals. As stated in the main text,
GATS Article I excludes services supplied in the exercise of government authority. In the domain of
audiovisuals, more than in that of other services, Governments can plausibly claim to have the right and the
obligation to protect and promote national culture, public morale and ultimately the supreme interests of the
country. They are therefore entitled to support domestic film productions or TV and radio programmes, for
instance, using instruments such as fiscal incentives, grants, soft loans and other financial subsidies, and public
enterprises. Article VIII allows, in general terms, the existence of monopolies and the establishment of new
ones. Thus, countries have ample policy room to establish new monopolies, or to re-establish previously
abolished ones, as can be the case for deregulated TV services. Under a state monopoly regime, any kind of
subsidy to public TV broadcasting would, of course, be quite acceptable. GATS Article XII allows government
procurement to foster the national production of audiovisual services. For instance, a government-owned TV
station might buy exclusively or preferentially domestically produced programmes, indirectly subsidizing
national producers. Still other GATS articles could be invoked by government to protect the national
audiovisual industry, not by means of subsidies but through more rudimentary policy tools. Article XIV
contemplates an exception to allow measures “necessary to protect public morals or to maintain public order”.
Article XIV bis states that: “Nothing in this agreement shall be construed…to prevent any member from taking
any action … for the maintenance of international peace and security”. Article XVII states (in a note) that
national treatment obligations should not be interpreted as a duty, on the part of members, to compensate for
“inherent competitive disadvantages which result from the foreign character of the relevant service or service
suppliers”. According to the European Union, this applies to situations where nationals prefer not to buy foreign
services for linguistic, cultural, religious, or for other reasons. Such an exception is likely to be applicable in a
particularly straightforward and broad fashion to audiovisual services, as demand for these services is strongly
determined by linguistic and cultural factors.
5.
SUBSIDIES IN THE AREA OF SERVICES IN DEVELOPING COUNTRIES:
A NECESSARY POLICY TOOL IN AN ASYMMETRIC WORLD
The content of the preceding sections can be summarized as follows. Section 1 shows that
Governments use several different policy instruments to implement subsidy programmes,
many of which are indirect in nature. Attempts to evaluate the magnitude of subsidies and, a
fortiori, estimate their sectoral, macroeconomic and international impact therefore face a
number of theoretical and practical difficulties. Section 2 argues that, although it has long
been recognized by economic theory that subsidies are frequently suboptimal, inefficient, and
even counterproductive, there are powerful factors belonging to the sphere of political
economy which tend to perpetuate and enlarge existing subsidy programmes and favour the
emergence of new ones. It also highlights the fact that subsidies for services, and for some
crucial service activities (such as R&D), can play a very important role as policy tools aimed
at increasing the productivity and competitiveness of national economies. Section 3 shows
that subsidies for services are huge, both in the North and in the South. However, due to
unequal financing capabilities and different policy priorities, there are major differences in
both the absolute and the per capita levels of subsidization of key service sectors between
developed and developing countries. Section 4 summarizes and interprets the evolution of the
38
treatment of subsidies over the various rounds of multilateral negotiations on international
trade. It concludes that the issue of subsidies remains controversial and, to some extent, an
ambiguous and unsettled one in the context of present multilateral trade diplomacy. It also
argues that GATS and other WTO agreements, so far, effectively shield services from most
of the binding norms limiting Government’s capacity to implement subsidy policies at will.
Taking stock of these findings, it is apparent that the question of subsidies acquires a very
different dimension in developed countries on the one hand and in developing countries on
the other. Moreover, the international community still does not sufficiently recognize this
asymmetry. In developed countries, subsidies are mainly used to pursue social, distributional
and environmental goals in the domestic realm and as strategic trade policy tools (aimed
primarily at overcoming foreign competitors of comparable technological strength) in the
international arena. In developing countries too, subsidies are used both for social and
economic purposes, but their role is at the same time more crucial and more delicate than in
developed countries. It is more crucial because the need for subsidies is far more pressing
than in developed countries, and few, if any, alternative policy tools are available to
developing countries’ policy-makers, due inter alia to relative institutional
underdevelopment. It is more delicate because in developing countries little money tends to
be available in government coffers (or even in citizens’ pockets), the whole economic
structure is fragile, information and data are scarce and not always reliable, and fine-tuning of
proactive interventions is difficult to achieve. In developing countries, therefore, wrong,
overstretched or simply poorly tailored and imperfectly targeted subsidy policies can cause
major distortions, severe financial crises, and long-lasting economic damage, much more so
than in developed countries. Still, according to many observers, practitioners and policy
makers, history shows that factors such as the very underdevelopment of market forces,
institutions and national private economic agents make exclusive reliance on laissez-faire a
very poor option for less advanced countries striving to achieve economic and social
development, to catch up progressively with the industrialized world and to consolidate
national sovereignty. According to this widely shared view, a certain level of interventionist,
proactive industrial policies, which usually entail the implementation of selective subsidy
policies as a key component, is a key ingredient of any development strategy, as it was, for
instance, for the United States, continental Europe and Japan in other historical periods
(Chang 2002). Governments in developing countries are therefore engaged in a very difficult
game—designing subsidy policies to support development goals while doing more with less,
i.e. to fulfil a more essential and strategic policy task than in the case of their counterparts in
the industrialized world, with fewer resources and less room for error.
The same fundamental argument applies to the use of subsidies for social rather than
development purposes. Subsidizing health, education, energy and other vital activities is far
more crucial for the fulfilment of basic human needs in the developing than in the developed
world, yet the resources available are meagre.
From a sectoral point of view, it is becoming increasingly clear that the emphasis of support
policies in developing countries, which might require the implementation of various types of
subsidies—be they earmarked for private or mixed agents or for governmental agencies
responsible for direct public provision—should be increasingly shifted towards services
rather than agricultural or industrial activities. Services, in fact, bear a large and still untapped
development potential in most developing countries, a potential that can directly bring about
important competitiveness and trade gains (as in the case of services such as tourism or of the
39
mushrooming new services based on advanced information technologies). Even more
importantly, the simultaneous achievement of social and economic objectives can result in
huge long-term gains for a country as a whole. The huge pay-off of early investment in
education in a country like the Republic of Korea, or the more modest but important
contribution of the traditionally prioritized health sector to Cuba’s balance of payments, are
only two among many examples. Moreover, in practical terms, the evolution of international
trade rules is expected to make it increasingly difficult to strengthen, or even maintain,
existing subsidy programmes in the domains of agriculture and industry (for both developing
and developed countries), while there is still plenty of room for subsidy policies in virtually
all services sectors. In this respect, it is important to ensure that the international community
confirm and implement the right of developing countries to be granted special, differential
and more favourable treatment to enact subsidy policies to support service activities. The
present high degree of freedom in utilizing subsidy policies in the domain of services should
be maintained only for developing countries, while the possibility of implementing subsidy
programmes should be progressively restricted in developed countries. Developed countries
should genuinely accept special and differential treatment as the building block of any
present and future negotiation on subsidies, not only on the basis of fairness and realistic
recognition of the weight of major structural differences between them and the developing
world, but also, consistent with the principle of enlightened and precautionary self-interest, to
avoid the risk of a gradual loss of credibility and acceptability for the multilateral trade
system as a whole.
40
REFERENCES
Aghion P and Howitt P (1998). Endogenous Growth Theory. Cambridge, MIT Press.
Benitah M (2001). The Law of Subsidies under the GATT/WTO System, Kluwer Law
international, The Hague, London, New York.
Bronwyn D et al. (2000). Is public R&D a complement or a substitute for private R&D? A
review of the econometric evidence. Research Policy, 29 (4-5):497-529.
Busch M (1999). Trade warriors – states, firms, and strategic–trade policy. In: Hightechnology Competition. Cambridge University Press, UK.
Busom I (1999). An Empirical evaluation of the Effects of R&D Subsidies, Economics of
Information and New technology, 9 (2):111-48.
Chang H-J (2002). Kicking Away the Ladder – Development Strategy in Historical
Perspective. Anthem Press.
Coase RH (1960). The Problem of Social Cost. In: Journal of Law and Economics (3).
De Moor A (1997). Perverse Incentives: Subsidies and Sustainable Development – Key
Issues and Reform Strategies. Earth Council Report. Institute for Research on Public
Expenditure, The Netherlands.
Eatwell J et al. (1987). The New Palgrave Dictionary of Economics (4).
International Energy Agency (IEA) (1999). Looking at energy subsidies: getting the prices
right. World Energy Outlook, Paris.
Gandolfo G (1998). International Trade Theory and Policy. Springer-Verlag, Berlin.
Griliches Zvi (1998). RD and Productivity. Series, National Bureau of Economic Research
Monograph
Guellec D and van Pottelsberghe de la Potterie (1997). Does government support/stimulate
private R&D? OECD Economic Studies No.29.
Grossman G and Helpman E (1991). Innovation and Growth in the Global Economy,
Chapters 11 and 12. MIT Press: Cambridge.
Lach S and Sauer R (2001). R&D, Subsidies and productivity, STE Working Paper, STE,
Israel.
Maddison D et al. (1997). Environmentally damaging subsidies in the UK. European
Environment (7):110-7.
Moore MR and McGuckin CA (1988). Reprint from US Department of Agriculture Report.
Program crop production and federal irrigation water, reprint Agricultural resources:
crop land, water, conservation situation and outlook USDA.
Morris RK. (1998). R&D Subsidies and Countervailing Measures, in S. Raymond, ed.,
Science Technology and the Economic Future, New York Academy of Sciences.
Myers N (1998). Perverse subsidies: taxes undercutting our economies and environments
alike. International Institute for Sustainable Development, Canada,
41
OECD
___
(1997). Reforming energy and transport subsidies – environmental and economic
implication, France.
___
(1998). Subsidy reform report. Improving the environment through reducing
subsidies, Paris.
___
(2001). Environmental outlook.
___
(2002). Science, technology and industry outlook.
O’Brien R (1997). Subsidy regulation and state transformation in North America, the GATT
and the EU, St. Martin’s Press, New York, XIV:210.
Repetto R (1986). Skimming the water: rent-seeking and performance in public irrigation
systems, research report no. 4, World Resources Institute, Washington DC.
Samuelson PA and Nordhaus WD (2001). Economics 17th edition, McGraw-Hll and Irwin.
SCM
___
(2002a). US Updating and New and Full Notification Pursuant to article XVI:1 of the
GATT 1994 and Article 25 of the SCM Agreement,
G/SCM/N/48/USA,G/SCM/N/60/USA, G/SCM/N/71/USA (2 July) WTO, Geneva.
___
(2002b). Ibid. Supplement, G/SCM/N/48/USA, G/SCM/N/60/USA,
G/SCM/N/71/USA/Suppl.1, (29 October) WTO, Geneva.
UNCTAD ___ (1993). The Impact of Subsidies on Trade in Services– Note by the UNCTAD
Secretariat, UNCTAD/SDD/SER/3, Geneva.
___ (1997). Report of the Expert Meeting on strengthening the capacity and expanding
exports of developing countries in the services sector: health services, TD/B/COM.1/1/7,
TD/B/COM.1/EM1.3. 16-18 June, Geneva .
___
(2000). Subsidies, Countervailing Measures and Developing Countries – with a focus
on the agreement on subsidies and countervailing measures, Report by the UNCTAD
Secretariat UNCTAD/DITC/COM/23.
___
(2000a). Report of the Expert Meeting on national experiences with regulation and
liberalization: examples in the construction services sector and its contribution to the
development of developing countries. TD/B/COM.1/32, TD/B/COM.1/EM.12/3. 23-25
October, Geneva.
___
(2001). Expert Meeting on Energy Services in International Trade: Development
Implications. Report. TD/B/COM.1/42, TD/B/COM.1/EM.16/3. 23-25 July, Geneva.
___
(2002). Report of the Expert Meeting on Audiovisual services: improving
Participation of Developing Countries, held at the Palais des Nations, Geneva, 13 to 15
November 2002, Doc. . TD/B/COM.1/56, TD/B/COM.1/EM.20/3.
Van Beers C and de Moor A (2001). Public Subsidies and Policy Failures, Edward Elgar
Publications, Cheltenham, UK.
42
Van Grasstek C (2001). Controversies over Cultural Trade: Hardware, Software, the Market,
and the State (Draft of June 23).
Winham GR (1986). International trade and the Tokyo round negotiations, Princeton,
Princeton University Press.
The World Bank
___
(1993). Water Resources Management. Washington DC.
___
(1994). World Development Report 1994. Washington DC.
WTO
___
(1998). Subsidies for Services Sectors – Information Contained in WTO Trade Policy
Reviews, Background Note by the Secretariat. S/WPGR/W/25, 26 January.
___
(2000a). Ibid. S/WPGR/W/25/Add.1, 29 May.
___
(2000b). Ibid. S/WPGR/W/25/Add.2, 12 December.
___
(2002). Ibid. S/WPGR/W/25/Add.3, 19 September.
43